DIIS WORKINGDIIS WORKING PAPER 20PAPER12:06

Continuity and change in ’s ruling coalition: legacies, crises and weak productive capacity Ole Therkildsen and France Bourgouin

DIIS Working Paper 2012:06 WORKING PAPER WORKING

1 DIIS WORKING PAPER 2012:06

OLE THERKILDSEN Senior Researcher, [email protected]

FRANCE BOURGOUIN Project Researcher, Ph.d. [email protected]

ACKNOWLEDGEMENTS The authors wish to express their gratitude to Goran Hyden, Per Tidemand, Rasmus Hundsbæk Pedersen and Sam Wangwe for their insightful comments and criticisms during the process of writing this paper. We also like to thank the members of the EPP team for their contributions. The usual disclaimers apply.

DIIS Working Papers make available DIIS researchers’ and DIIS project partners’ work in progress towards proper publishing. They may include important documentation which is not necessarily published elsewhere. DIIS Working Papers are published under the responsibility of the author alone. DIIS Working Papers should not be quoted without the express permission of the author.

DIIS WORKING PAPER 2012:06 © The authors and DIIS, Copenhagen 2012 Danish Institute for International Studies, DIIS Strandgade 56, DK-1401 Copenhagen, Denmark Ph: +45 32 69 87 87 Fax: +45 32 69 87 00 E-mail: [email protected] Web: www.diis.dk Cover Design: Carsten Schiøler Layout: Allan Lind Jørgensen Printed in Denmark by Vesterkopi AS ISBN: 978-87-7605-490-8 Price: DKK 25.00 (VAT included) DIIS publications can be downloaded free of charge from www.diis.dk

2 DIIS WORKING PAPER 2012:06

DIIS WORKING PAPER SUB-SERIES ON ELITES, PRODUCTION AND POVERTY

This working paper sub-series includes papers generated in relation to the research programme ‘Elites, Production and Poverty’. This collaborative research programme, launched in 2008, brings together research institutions and universities in Bangladesh, Denmark, Ghana, Mozam- bique, Tanzania and and is funded by the Danish Consultative Research Committee for Development Research. The Elites programme is coordinated by the Danish Institute for International Studies, Copenhagen, and runs until the end of 2011.

Earlier papers in this subseries: Rweyemamu, Dennis: “Strategies for Growth and Poverty Reduction: Has Tanzania’s Second PRSP Influenced implementation?” DIIS Working Paper 2009:13. Kjaer, Anne Mette, and Fred Muhumuza: “The New Poverty Agenda in Uganda” DIIS Work- ing Paper 2009:14. Whitfield, Lindsay: “The new ‘New Powerty Agenda’ in Ghana: what impact?” DIIS Working Paper 2009:15. Webster, Neil, Zarina Rahman Khan, Abu Hossain Muhammad Ahsan, Akhter Hussain and Mahbubur Rahman: “State Elites and the New Poverty Agenda in Bangladesh”. DIIS Work- ing Paper 2009:22. Buur, Lars, with Obede Suarte Baloi: “The Mozambican PRSP Initiative: Moorings, usage and future” . DIIS Working Paper 2009:35. Whitfield, Lindsay: “Developing Technological Capabilities in Agro-Industry: Ghana’s ex- perience with fresh pineapple exports in comparative perspective”. DIIS Working Paper 2010:28. Whitfield, Lindsay: “How countries become rich and reduce poverty: A review of heterodox explanations of economic development”. DIIS Working Paper 2011:13. Whitfield, Lindsay and Ole Therkildsen: “What Drives States to Support the Development of Productive Sectors?”, DIIS Working Paper 2011:15. Buur, Lars and Lindsay Whitfield: “Engaging in productive sector development: Comparisons between Mozambique and Ghana,” DIIS Working Paper 2011:22. Whitfield, Lindsay: “Competitive Clientelism, Easy Financing and Weak Capitalists: The Con- temporary Political Settlement in Ghana”, DIIS Working Paper 2011:27. Whitfield, Lindsay: “Growth without Economic Transformation: Economic Impacts of Gha- na’s Political Settlement”, DIIS Working Paper 2011:28. Whitfield, Lindsay: “Political Challenges to Developing Non-Traditional Exports in Ghana: The Case of Horticulture Exports”, DIIS Working Paper 2011:29. Buur, Lars with Obede Baloi and Carlota Mondlane Tembe: “Mozambique Synthesis Analysis: Between Pockets of Efficiency and Elite Capture,” DIIS Working Paper 2012:01. Kjær, Anne Mette, Fred Muhumuza and Tom Mwebaze: “Coalition-driven initiatives in the Ugandan dairy sector: Elites, conflict, and bargaining”, DIIS Working Paper 2012:02. Kjær, Anne Mette, Fred Muhumuza, Tom Mwebaze, Mesharch Katusiimeh: “The political economy of the fisheries sector in Uganda: ruling elites, implementation costs and industry interests”, DIIS Working Paper 2012:04. Kjær, Anne Mette, Mesharch Katusiimeh, Tom Mwebaze and Fred Muhumuza: “When do rul- ing elites support productive sectors? Explaining policy initiatives in the fisheries and dairy sectors in Uganda”, DIIS Working Paper 2012:05. Kjær, Anne Mette and Mesharch Katusiimeh: “Growing but not transforming: Fragmented ruling coalitions and economic developments in Uganda”, DIIS Working Paper 2012:07.

More information about the research and access to publications can be found on the website www.diis.dk/EPP or via the Q2 barcode

3 DIIS WORKING PAPER 2012:06

4 DIIS WORKING PAPER 2012:06

CONTENTS

Abstract 7 1. Introduction 9 2. Present structure of the ruling coalition: party, bureaucracy, military 12 2.1 The CCM and its dominance 13 2.2 The bureaucracy 15 2.3 The military 16 2.4 Weak entrepreneurs 16 2.5 Conclusions 18 3. The subtle causes of coalition changes 18 3.1 Structural changes in the Tanzanian economy 18 3.2 The importance of education for elite formation 20 3.3 Conclusion 21 4. Crisis and changes in Tanzania’s ruling coalition over time 21 4.1 The roots of CCM dominance and its changes over time 22 4.2 The Bureaucracy 27 4.3 The Military 30 4.4 Emerging entrepreneurs 30 4.5 Conclusions 37 5. Resource use to stay in power 38 5.1 Taxes and aid 38 5.2 The Party 39 5.3 The bureaucracy 42 5.4 The military 44 5.5 Emerging entrepreneurs 44 5.6 Conclusions 47 6 Political stability without sustained economic development 47 References 50

5 DIIS WORKING PAPER 2012:06

LIST OF ABBREVIATIONS

ACT Agricultural Council of Tanzania ASM Artisanal-scale mining BIS Basic Industry Strategy BoT Bank of Tanzania CC Central Committee CCM CHADEMA Chama cha Maendeleo na Demokrasia CSO Civil society organisation CTI Confederation of Tanzanian Industries CUF Civil United Front DRD Domestic Revenue Department EPA External Payments Account EPP Elites, Production and Poverty research programme FDI foreign direct investments GDP Gross domestic product Golkar Partai Golongan Karya (Indonesia) IMF International Monetary Fund LSM Large-scale mining LTD Large taxpayer department MP Member of Parliament MVIWATA National Network of Smallholder Farmers NEC National Executive Committee NGO Non-governmental organization OECD Organisation for Economic Co-operation and Development SAP Structural Adjustment Programmes SOE State-owned enterprises STAMICO State Mining Corporation TANU Tanganyika African National Union TCCIA Tanzania Chamber of Commerce, Industry and Agriculture TCME Tanzania Chamber of Mines and Energy TNBC Tanzania National Business Council TPSF Tanzania Private Sector Foundation TRA Tanzania Revenue Authority

6 DIIS WORKING PAPER 2012:06

ABSTRACT

This paper presents analyses of the current composition of Tanzania’s ruling coalition, comprised mainly of the ruling party (the CCM), the bureaucracy and the military, of how it has changed over time and of how its funding has evolved. Specifically, it discusses how historical legacies, structural changes in the economy and specific crises have influenced the composition of the ruling coalition, the holding power of its factions and the strategic use of resources to maintain its power. The paper concludes that Tanzania’s ruling coalition is presently characterised by conflicts and bargaining among strong factional elites within the ruling coalition and by the increasing power of its lower level factions. Opposition parties are largely excluded from influence and remain weak. Economic entrepreneurs in the formal productive sectors are few and poorly organised. Their relations with the ruling coalition are ambiguous and largely informal, although exchanges of money and rents are of increasing importance in the relationship. Moreover, informal sector entrepreneurs and smallholders in agriculture are largely excluded from the ruling coalition. There is little evidence that the ruling coalition – despite decades of political stabil- ity – has used its position to build and strengthen the productive capacity of domestic entrepreneurs.

7 DIIS WORKING PAPER 2012:06

8 DIIS WORKING PAPER 2012:06

1. INTRODUCTION spite impressive economic growth over the past fifteen years (Nord et al. 2009, Figure The present ruling party in Tanzania, the 1). Furthermore, Tanzania’s headcount pov- CCM (and its predecessor, TANU), has been erty declined by only 2.4 percentage points in power since the country’s independence in between 1991 and 2007, compared, for ex- 1961, longer than any other political party on ample, to more than 20 percentage points in the continent (Smith 2005, table 1). Moreo- Ghana and Uganda (Hoogeveen and Ruhin- ver, the key features of Tanzania’s ruling duka 2009, 2; Treichel 2005, 17). The actual coalition until the early 1990s were ‘close to number of poor people is, in fact, increasing that of a developmental state’ (Khan 2010, yearly (Hoogeveen and Ruhinduka, 2009, 64), 120). Groups that could potentially threaten although there may have been advances in the power of the ruling coalition were almost other aspects of peoples’ lives (such as educa- non-existent, while internal lower-level fac- tion and health). Yet, measured by experienc- tions remained weak. Since 1995 the CCM es of ‘lived poverty (how frequently people has also won all competitive elections at the go without basic necessities during the course national and local levels by significant mar- of a year),’ Tanzanians actually saw the qual- gins. Consequently, the ruling coalition has ity of their daily lives worsen between 2000 had a relatively high degree of stability for and 2005 (Mattes 2008, 2 & 16). fifty years. This makes the case of Tanzania Even this lacklustre economic perform- unusual in Africa. ance since independence, especially during Yet despite this political stability, produc- the hardship years of the 1980s, has not lead tive sector capacity has not improved sig- to the emergence of a strong political opposi- nificantly since independence. Productivity, tion, nor have there been any sustained popu- especially in agriculture, has remained largely lar protests against structural adjustment, as stagnant for fifty years (Skarstein 2010). And has been the case elsewhere on the continent while Tanzania’s growth has accelerated in (Sandbrook 2000, 85). Indeed, the CCM has recent years, particularly in gold, manufactur- won all competitive elections since 1995 by ing goods, tourism and services, it has largely substantial margins, except on Zanzibar. And been driven by foreign direct investments while many prominent political leaders have (FDI) and aid, which have not created many been implicated in large-scale corruption jobs nor led to significant positive structur- scandals in recent years (Anti-Corruption al changes (Research and Analysis Working Resource Centre 2009), the legitimacy of the Group 2009). Indeed, like in other African party does not yet seem to have been signifi- and Latin American countries, the structural cantly compromised. Almost three-quarters changes that have occurred in Tanzania may of adult Tanzanians were reported to ‘feel well have resulted in labour moving ‘in the close’ to the CCM in 2008 (Afrobarometer wrong direction, from more productive to less and REPOA 2008), and the 2010 elections productive activities, including, most notably, were again won by large margins, repeating informality’ (McMillan and Rodrik 2011, 1- the pattern from the three previous competi- 3). tive elections since 1995. The bottom line is that living standards for The case of Tanzania therefore illustrates ordinary people measured by GDP per capita that political stability is far from sufficient have not improved much since the 1970s, de- to ensure sustainable economic growth and

9 DIIS WORKING PAPER 2012:06

poverty alleviation. Ruling elites must also be choose policies and implementation arrange- motivated to promote structural change and ments as part of their strategies for main- technological upgrading through public in- taining power,3 in particular for maintaining vestments in activities with short-term risks ruling coalitions4 and/or winning elections’; but uncertain, long-term development ben- 3) ‘good economic outcomes depend on a) efits, and by public promotion of technology close relations between the ruling elite and acquisition and learning that compensate for the relevant productive entrepreneurs based uncompetitive cost structures while learning on mutual interests between them; and b) the takes place (Whitfield and Therkildsen 2011, ability of the ruling elite to create pockets of 8). Or, to paraphrase Gray (2011, 259), while bureaucratic capabilities to implement spe- political stability is clearly important, rul- cific policies’ (see Whitfield and Therkildsen ing elites must also promote the two other 2011). important drivers of structural change to The characteristics of the ruling coalition achieve sustainable growth and poverty alle- – its vulnerability, degree of fragmentation viation: primitive accumulation and technol- and sources of finance – are shaped by the ogy acquisition. country-specific structural and institutional This paper presents an analysis of the setting, and by the strategies used by the rul- Tanzanian ruling coalition that highlights the ing elite to stay in power. Crises can have a tensions in power relations and the potential particular impact on the vulnerability and threats to the hegemony of the ruling party fragmentation of ruling coalitions. Systemic since independence. This will help us under- crises, which threaten the survival of a rul- stand why Tanzania’s ruling elite has not been ing coalition, can create ‘pressure to mitigate able to use the state and its resources to pro- the inimical economic consequences of [a] mote structural change and achieve sustain- fragmented ruling elite. The number of veto able poverty alleviation in a decisive manner, players is “reduced” either by a formal reduc- despite its long-term and long-enunciated vi- tion of the actual number of veto players (via sion to ‘transform agriculture’ and become a constitutional change, coups, party restruc- ‘semi-industrial country’ by 2025 (Rweyema- turing, elections), by aligning preferences and mu 2009, 14; McHenry 1994). thus reducing the effective number of veto The starting point for such analyses is the players, and by relying on delegation of poli- conceptual framework developed by the Elit- cymaking in order to “tie the hands” of the es, Production and Poverty (EPP) research ruling elite in order to increase decisiveness programme at DIIS.1 It makes three propo- of policymaking’ (Whitfield and Therkildsen sitions about the incentives driving policy 2011, 23). Moreover, economic crises ‘can choices made by ruling elites: 1) ‘ruling elites2 support the development of productive sec- 3 The holding power of a faction inside or outside the ruling tors when they perceive that this will help coalition is based on its economic wealth, legitimacy and its them to remain in power’; 2) ‘ruling elites ability to organise supporters and to mobilise them politi- cally. 4 The ruling coalition refers to the ruling elite, as well as the 1 See http://www.diis.dk/ for more information on the EPP groups and individuals who support the rise and survival programme. of the ruling elite, typically in exchange for benefits. These 2 The ruling elite is the group of people who wields formal or groups and individuals keep the ruling elite in power by or- informal power to make or influence authoritative decisions ganizing political support, often in the form of patron-client in government and/or the party in power. networks.

10 DIIS WORKING PAPER 2012:06

change the characteristics of the ruling coa- legitimacy crisis in the early 1990s, combined lition by shrinking state revenue available to with donor pressure for democratisation, led ruling elites for building coalitions and win- to the (re)introduction of competitive elec- ning elections, by reducing foreign exchange tions. This has helped shift power towards earnings, and by drying up political funds. lower level factions within the coalition and The resulting lack of easy access to resources increased the power of entrepreneurs whose can affect ruling elites’ motivations, and how financial support is needed to win elections they engage with productive entrepreneurs’ and build coalitions. Finally, growing resourc- (Whitfield and Therkildsen 2011, 26). es since the mid-1990s (revenues, aid, rents) Using key elements of this conceptual are increasingly being used by the ruling coa- framework, we find first that since the early lition to build and maintain itself and to win 1960s Tanzania’s ruling coalition has con- elections. sisted of overlapping factions of the CCM, Thus, what emerges is a political settlement the bureaucracy and the military. Only fairly which – over the last fifty years – has been recently – and reflecting the increasing im- remarkably stable, but has also changed in portance of money in Tanzanian politics many respects. However, these changes have – have some economic entrepreneurs gained not led to significant structural changes in the sufficient holding power to influence policy economy, and there is little evidence that the making and implementation, mostly through resources controlled by the ruling coalition are informal relations to the ruling elite. Sec- systematically being used to build capacity in ondly, over time lower level factions of the the productive sectors. There are three main party and the bureaucracy have gained power, explanations for this. The reintroduction of but elite bargaining continues to be centred competitive elections since 1995 has not re- on factions within the coalition, which is held duced clientelistic politics. On the contrary, together by the CCM’s formidable ability to it has motivated the ruling elite to pursue win competitive elections, although elections populist rather than growth-enhancing poli- in Zanzibar are an exception to this. Thirdly, cies.5 Moreover, elections have aggravated the historical legacies have shaped the present internal fragmentation of the ruling coalition structure of the ruling coalition, as first the and increased competition among its factions. colonial rulers and subsequently the one-par- At the same time, the lower level factions of ty state motivated by economic nationalism the coalition have gained in strength (Khan have tried to suppress indigenous capitalist (2005)). Consequently, it has become difficult entrepreneurs. Still today these groups re- to establish pockets of administrative effi- main weak. Fourthly, over time major crises ciency or to shift state resources that could have shaped the coalition. The mutiny crisis in help to build capacity in the productive sec- 1964 jolted the military into the ruling coa- tors. Finally, a viable political settlement be- lition, where it continues to play an impor- tween the CCM and mostly Asian-Tanzanian tant role. The Structural Adjustment Programmes entrepreneurs, which dominates the produc- (SAP) crisis resulted from the economic tur- tive sectors in the formal economy (alongside moil of the 1980s,leading to a gradual and foreign firms), has not yet emerged due to the uneven liberalisation of the economy, which weakened the ruling coalition and enabled 5 Kjaer and Therkildsen (2012) have explained some of the more entrepreneurs to emerge. The CCM’s motivations of ruling elites to make such decisions.

11 DIIS WORKING PAPER 2012:06

ruling party’s ideology of economic nation- for its power. Current relations with produc- alism. This economic nationalism, which has tive entrepreneurs are individualistic and am- broad public backing (Peter 2002), has ham- biguous, although money politics is on the pered the design and implementation of in- rise. Excluded factions in the opposition are dustrial policies which are the key to sustain- weak. able growth and poverty alleviation in poor Other scholars have also conducted analy- countries (Cimoli, Dosi, and Stiglitz 2009). ses of power for Tanzania, which put the The paper is organised into five parts. It findings of this paper in perspective. Khan begins by presenting a snapshot overview of (2010), for example, argues that Tanzania’s the present composition of the Tanzanian rul- ruling coalition is centred on a weak dominant ing coalition in Section 2. The following three party. However, he overlooks the fact that sections then trace various aspects of change the Tanzanian coalition differs from ruling in the coalition over time. Section 3 addresses coalitions in other countries in that factional two subtle causes of coalition change: struc- conflicts about power occur within the up- tural changes in the economy and education. per echelons of the party, as well as between Section 4 analyses changes in power relations higher and lower level factions within the rul- within the coalition in response to the three ing coalition. For these reasons, we argue that crises, showing how these have affected the the ruling coalition is better characterised in emergence of an entrepreneurial class. Sec- terms of its competitive clientelism between tion 5 explains how resources are generated many strong internal factions. Using different and used by the ruling coalition specifically to conceptual frameworks than EPP, analyses by build and sustain itself and to win elections. Cooksey and Kelsall (2011) arrive at similar The final section of the paper, Section 6, general conclusions in most respects, as does presents conclusions and summarises the an- Wangwe (2010). Khan and Gray (2006, 50) swers to why Tanzanian’s ruling coalition has and Lindemann and Putzel (2008, 2) empha- been unable to spearhead structural changes sise the inclusive nature of elite bargaining in the economy that could lead to sustained in Tanzania. The latter argue that the ruling growth and poverty alleviation. party ‘has been able to forge and maintain a centralised and inclusive political coalition’ in which the organisation of patronage is ‘cen- 2. PRESENT STRUCTURE OF tralised’ and elite bargaining is ‘inclusive of all THE RULING COALITION: significant sources of political, military, eco- PARTY, BUREAUCRACY, MILITARY nomic and social power within the territory of the Tanzanian state.’ However, they overlook Political power in Tanzania today is mainly the fact that the opposition parties have been centred around the ruling party, the CCM, largely excluded from influence (only recently and in particular its central organs and the did the CCM enter a power-sharing arrange- president (who is also the chairman of the ment with the opposition on Zanzibar), and party). Overlapping factions of the CCM, they also ignore the often coercive nature of along with factions in the bureaucracy and the state’s relations with civil society associa- military, make up the ruling coalition. This tions (Mercer 1999). integration of factions is a key feature of the Similarly, our analysis also differs some- ruling coalition and an important explanation what from other analyses of power relations

12 DIIS WORKING PAPER 2012:06

in Tanzania. For instance, McHenry (1994) The present structure of the ruling coalition argues that ideological conflicts between ide- is the result of historical legacies in education alist and pragmatic socialists have been cen- and structural changes in the economy, as ex- tral to power struggles – a view that Ewald plained in section 3, as well as of the effects (2011, 340-343) also seems to support. But of various crises, as explained in section 4. we argue that today real conflicts are main- But first a snapshot of the present ruling coa- ly about access to resources and about how lition is presented. to win elections to stay in power. In their analysis of the Tanzanian parliament’s role in the first PRSP process at the end of the 2.1 The CCM and its dominance 1990s, Gould and Ojanen (2003) argue that The power of the ruling party is based on this policy process was dominated by an iron three complementary factors: (a) a relatively triangle, excluding the CCM, of donor and strong and institutionalised party organisation state technocrats working closely with a select that reaches down to the lowest level of soci- group of ‘constructive’ NGOs. Although this ety, giving the party significant advantages in is plausible, it ignores the fact that the PRSP competitive elections and in the control and is not a good indicator of the political priori- use of rents; (b) close links to the bureauc- ties of the ruling elites. Indeed, the PRSP was racy and the military – two of the largest and mainly prepared to attract donor funds and best organised entities in Tanzania – through qualify for debt relief and has not been a reli- which inclusion and control over all signifi- able guide to what was actually implemented cant sources of political, economic and social on the ground (Rweyemamu 2009). power is exercised; and (c) the continuing le- Hyden (2005) and Mmuya and Tostesen gitimacy of the CCM. (2007) regard donors as the most powerful A strong organisation exists within the actors in policy design and implementation. party. Power is centred in the party chairman, Moreover, the latter argue that ‘centralisation who is also the Union president, in the 30+ remains a main feature of Tanzania’s power member Central Committee (CC) and in the structure.’ This is an exaggeration. Based on National Executive Committee (NEC). A the analyses presented below, the position on national level secretariat gives administrative donors taken by Lawson and Rakner (2005) and technical muscle to the party. It prepares is more plausible: although donor influence the election manifestos and other party docu- is tolerated (but often deeply resented), ‘in ments that have some influence on govern- practice it is domestic political factors which ment policies and thinking; it helps to co- exert a dominant influence over policy’ and ordinate the regional and district level party its implementation. It is, however, interesting organs; it distributes party funds to these or- that the authors mentioned above support gans; it helps the party to monitor govern- our analyses with respect to the power of en- ment activities; etc. trepreneurial groups. These are broadly seen During the last elections in 2010, the CCM as having limited although growing influence won 258 of the 343 seats in parliament, and on policy making and implementation. Only a 63 per cent of the presidential vote. The mar- few entrepreneurs are openly part of the rul- gin of victory was less than in the elections of ing coalition, although some have close per- 2000 and 2005, but no worse than the results sonal relations with factions within it. of the 1995 elections. The opposition does

13 DIIS WORKING PAPER 2012:06

not (yet) pose a serious threat to the CCM’s ent from that of the CCM. However, we do hegemony – except on Zanzibar – and has, know that civil society in Tanzania is gener- perhaps, helped to reinvigorate the CCM, as ally weak despite the growth in NGOs since shown below. At district and village level the the 1990s.9 Generally, these organisations are CCM dominance is even higher (see Section still closely monitored or controlled by the 4.1.3). party and/or government, including the local The ability of the party to reach the grass- administration (Mercer 1999). roots across the country is a result of the cen- Nevertheless, factional conflicts within the tralised power within the party and its delib- party leadership abound. Some have run for erate party- and coalition-building activities. years (see Section 4). More recently, in the af- Consequently, the social base of the CCM termath of the 2010 elections, conflicts over today is very wide and encompasses mem- the use and access to rents and the control of bers from many different state and non-state the party have exploded. These are the reac- elites, including members of the police, army tion of President Kikwete and his group to and state intelligence. Its geographical reach the relatively poor CCM performance in the is also wide (Mmuya, 1998, 33-35). The par- 2010 elections and to the threats that some ty’s social base therefore cuts across regional, CCM leaders pose to his grip on the party. ethnic and social divides. It claims to have 3.5 Thus, on 9 April 2011, in an unprecedented million members, far surpassing the member- move, the Central Committee – including the ship of other parties (Ewald 2011, Appen- whole of its secretariat and its Secretary Gen- dix 1). None of the opposition, according to eral, Yusuf Makamba10 – resigned, and Kik- Mukandala et al. (2005, 5), ‘has cultivated or wete as party chairman was asked to form a entrenched itself in significant “social con- new team to achieve ‘a real transformation of stituencies” nor with the mushrooming civil the party, to rectify its low morale and dis- society organisations.’ This may be an exag- cipline’.11 He appointed new members and geration. CUF clearly draws its support main- dropped some powerful old ones, such as ly from Zanzibar and from the Muslim popu- the former Prime Minister and Monduli MP lation.6 Some of CHADEMA’s new members Edward Lowasa (implicated in the Richmond of parliament started out as NGO activists scam,12 which forced him to resign as PM in on land issues.7 And in the 2010 elections the 2008), Igunga MP (past head opposition did relatively better in urban con- of party finance and a prominent business- stituencies and in the regions around Lake 8 Victoria and Kilimanjaro. Unfortunately, 9 Figures differ. Ewald (2002, 8) claims the existence of 9000 apart from such observations, we do not NGOs in 1999, while 4000 registered CSOs/NGOs are men- tioned by Mukandala et al. (2005, viii). know enough to decide whether the social 10 His son, , was appointed a member of base of the opposition is significantly differ- the CC at the same time – perhaps as compensation for his father’s expulsion? He takes up the post of head of foreign affairs, replacing the minister for Foreign Affairs and Interna- 6 And, for the first time ever, the CUF has won a couple of tional Co-operation. Makamba junior is also the chairman of seats on the mainland in the 2010 elections. the parliamentary committee on energy and minerals. ‘Fresh hope as the CCM announces teams.’ The Citizen, 12 April 7 CHADEMA is the largest opposition party. ‘Land conflicts 2011. shift political alignments’, The Citizen, 14 November 2010. 11 Tanzanian Affairs. http://www.tzaffairs.org/2011/05/the-re- 8 The opposition to TANU/CCM in the 1960s also came newal-of-ccm/ (accessed 1 June 2011). from these regions, a pattern that may be re-emerging now (Erdmann (2007, 20)). 12 See Anti-Corruption Resource Centre (2009).

14 DIIS WORKING PAPER 2012:06

man thought to contribute substantial funds The third factor that helps explain the CCM’s for the party) and Bariadi West MP Andrew power is its continuing high degree of legiti- Chenge (previous Attorney General impli- macy. It is remarkable that Tanzanians are cated in the BAe radar purchase scandal, but relatively positive in their assessment of the with substantial regional support).13 government’s overall performance, regard- In addition to these national level conflicts, less of their personal current living standards lower level factions of the party have gained – and this despite the fact that sustained eco- power. As already mentioned, the CCM domi- nomic growth in the 2000s has not had much nates the lower level political bodies (district impact on poverty alleviation and has not and village councils), which are increasingly significantly improved incomes for the great important for its mobilization of votes to majority of people. Corruption scandals im- win elections. In addition, CCM candidates plicating top CCM and government officials for parliament are nominated by the lower do not yet seem to have affected the party’s levels of the party – subject to screening by reputation significantly either (Mattes 2008). higher level organs, and assisted by members of the intelligence service – before they can contest the elections as party candidates . 2.2 The bureaucracy Hitherto, only party leaders at the district and The party has for many years maintained a village levels have participated in this process. firm grip on the bureaucracy, which now em- The CCM, however, changed its nomination ploys more than 300,000 people. Access to rules just prior to the 2010 elections so that powerful positions in the bureaucracy is con- all party members were given the right to vote trolled by the president, who also chairs the in the nomination process. Consequently, it party. He has extensive influence on these became more costly for candidates seeking a matters15 (in addition to his more wide-rang- nomination to pay bribes, as more individu- ing powers to issue directives). Although als with voting rights would now have to be some of these powers have recently been del- paid. This change in the nomination proce- egated, this does not preclude the president dure was pushed by Kikwete’s supporters in himself from taking decisions on personnel an effort to reduce the power of wealthy poli- matters if he so desires. Practised for decades ticians (such as Lowasa, Aziz and Chenge) in (Bana & McCourt, 2006), these powers have buying support for their protégés during party helped intertwine the CCM and government, nomination campaigns. However, a few days an important basis of the CCM’s influence. before the party nomination, Conversely, the bureaucracy has substantial removed the identity card control requirement holding power vis-à-vis the party. The impor- designed to ensure that only genuine party tance of the bureaucracy for the CCM has in- members be allowed to vote. Substantial num- bers of counterfeit membership cards were 15 Specifically, the president appoints, promotes, terminates, revokes appointments, transfers and dismisses the Chief Sec- produced, and this undermined the intended retary (who is also head of the civil service and Secretary purpose of the new nomination rule.14 to the Cabinet); Permanent Secretaries and their deputies; Heads of extra-ministerial/independent departments; Region- al Administrative Secretaries for Regions; Regional and Dis- trict Commissioners; High Commissioners and Ambassadors; 13 Section 5 presents more detail about the rents. members of the Public Service Commission; etc. (Bana and 14 Interview with NEC member (January 2011) and confirmed McCourt 2006, table 1). Divisional Secretaries with powers of in additional interviews. arrest are also presidential appointees.

15 DIIS WORKING PAPER 2012:06

creased since the introduction of multiparty tary staff has been integrated into the party elections, as the loyalty of public employees is and the bureaucracy. This has meant that crucial to maintaining a majority vote to win the army officers could become cabinet min- election. Members of staff from the regional isters either through election to parliament administration, as well as district and village or by direct presidential appointment – and council employees, are especially important they still can.16 The general presence of the for mobilizing votes and helping the party to military in the party and the bureaucracy has run its campaigns. Since grassroots elections grown significantly. This is illustrated by the are held one year prior to local government following hard to obtain (and therefore scat- and national elections, the CCM members in tered) information from Lindemann and Put- political offices at that level are used to mobi- zel (2008) and from our own information. In lise voters for the subsequent higher level elec- 1987, 15 percent of the CCM’s Central Com- tions (Tidemand and Msani 2008). Moreover, mittee members had a security/military serv- government staff are among the best organ- ice background, a percentage that increased to ised in the country and therefore constitute 20 percent by 2010.17 In 1982, more than 30 important interest groups in their own right. percent of all regional party secretaries were It is, for example, mainly the Teachers’ Union military personnel, though by 1987 this figure (the largest in the country) which has threat- had fallen slightly to 24 percent. Similarly, in ened to strike in recent years. 1982 some 15 percent of District Commis- Factional conflicts, however, reduce the sioners were army officers. This percentage power of the bureaucracy. Inter-ministe- rose to 20 percent in 2000 and is now around rial conflicts over their respective mandates, 25 percent.18 Moreover, intelligence officers, ministerial power vis-à-vis local authorities whether in post or retired, are actively in- and their access to donor funds have run for volved in the screening and vetting of CCM years and are still prominent features of to- candidates for various posts in the party and day’s public sector (Therkildsen 2000b). This in the nomination of party candidates for seriously affects the coordination of govern- the local government and parliamentary elec- ment initiatives between ministries, as well as tions. This, of course, gives them substantial between these and local governments. Neither influence. the Cabinet nor the Cabinet Secretariat – its administrative arm – have been able to change this significantly (Mukandala et al. 2008). 2.4 Weak entrepreneurs Some of these coordination problems have Some Tanzanian entrepreneurs maintain been exacerbated since the decentralisation close, albeit mostly informal relations with reform started in the late 1990s giving lower members of the ruling coalition, but they level factions of the bureaucracy substantially are not (yet) a prominent part of it – nor are more power, as discussed in Section 4.2. they many. Black Tanzanian capitalist entre-

16 2.3 The military Indeed, both the current president and prime minister worked in the intelligence services before taking up their As will be explained in Section 4.3, the mili- present positions. tary mutiny in 1964 jolted army staff into to 17 2009 information from own data base. the ruling coalition. Over the years the mili- 18 The recent figure is from own data base.

16 DIIS WORKING PAPER 2012:06

preneurs are relatively few and far between 2002)19 and still presents a political problem and are only now emerging. Productive sec- for the ruling coalition (Gray 2011). In ad- tor companies operating in Tanzania’s for- dition, it seems that good relations with the mal economy are still mostly foreign-owned, CCM are an important factor for success in usually by South African entrepreneurs, or, business. Thus on 7 April 2000, Makidara if the owners are Tanzanian, are typically of Mosi, a member of parliament for the op- South Asian or Arab descent. According to position party NCCR, requested clarification Cooksey and Kelsall (2011, 52) this is also from the Prime Minister, Frederick Sumaye, the case in most other large commercial who had said that, ‘If you want to do well in enterprises found in the transport, import/ business you should join the CCM.’ Sumaye export, retail, tourism, telecommunications, denied this implication, but it is a fairly wide- urban real estate and media sectors. spread view among many business people Only a few black Tanzanian businessmen interviewed in the late 2000s that connec- have been able to compete with non-black tions with the CCM can be very helpful for Tanzanian or foreign-owned businesses. business (Babeiya 2011, 95). Cooksey and Kensall refer specifically to The poor organisation of Tanzanian en- the media and soft drinks mogul Reginald trepreneurs in the productive sectors only Mengi as Tanzania’s only black African ‘ty- further limits their holding power vis-à-vis coon’ (he was also the past chairman of the the ruling elite. The case of the Tanzania Pri- Confederation of Tanzanian Industries or vate Sector Foundation (TPSF) in this regard CTI). He is currently using his newspapers is instructive. Established in 1998 with sup- to promote the government’s ‘Agriculture port from the World Bank and other donors First’ initiative. Other significant players (who still provide the bulk of its funding), it include Iddi Simba (interests in manufac- is a self-declared apex of business associa- turing, trade and, agriculture; former Min- tions. Together with the Tanzanian National ister of Industry; former chairman of CTI), Business Council – another donor-supported Juma Mwapachu (interests in investment forum for government–business dialogue and consultancy services, travel agencies – it is the government’s preferred association and shipping; former chairman of CTI, when dealing with the private sector. Observ- former ambassador and former Secretary ers argue that the TPSF has been co-opted General of the ) by the government. Its present head, Esther and Ali Mfuruki (interests in soft-ware and Mwikuza, has very limited business experi- retail, and a former member of the board ence but used to run an NGO which had the of the Bank of Tanzania). Black Tanzanian previous president’s wife as a patron. How- businessmen such as these have close con- ever, the TPSF has been weakened recently tacts to the party, and some help finance its by the withdrawal of three important mem- election campaigns. bers (CTI, Tanzania Chamber of Mines and The weak position of non-black entre- Energy (TCME) and the Tourism Confedera- preneurs is fuelled by a general distrust to- tion of Tanzania) who complained that the wards them. Economic nationalism is still very much alive. The so-called indigenisa- 19 A recent PhD by Matiku (2008) shows that economic na- tionalism also has strong supporters in the academic com- tion debate intensified after the privatisation munity, although there are also robust opposing views (Peter of parastatals began (Aminzade 2003; Peter 2002).

17 DIIS WORKING PAPER 2012:06

TPSF excludes them from influence and that zation with an infrastructure unmatched by it is not effective in negotiating with govern- any other party, and it seems likely to remain ment.20 Most members of the TPSF board dominant for at least another decade’. The do, in fact, represent smaller traders through election winning capabilities of the the CCM their membership in the Tanzania Chamber machinery is still so formidable, that political of Commerce, Industry and Agriculture (TC- entrepreneurs stand a much better chance of CIA). The Confederation of Tanzanian In- being elected as candidates of the CCM than dustries (CTI), established in the early 1990s, outside it. This prediction is stated with some has many large companies as members, but caution as it is not yet known if the recently its chairman has always been a strong CCM purged powerful the CCM Central Commit- supporter (see Section 4.4.2). tee members will react to their expulsion by starting a new party.

2.5 Conclusions Tanzania’s ruling coalition presently consists 3. THE SUBTLE CAUSES OF of factions of the party, the bureaucracy and COALITION CHANGES the military. The power of productive entre- preneurs is growing but remains relatively The current structure of Tanzania’s ruling weak, while the informal links between im- coalition described above has come about port/export traders and commercial business- in part as a result of structural changes in people with the ruling coalition are stronger the economy over time. The historical lega- than those of productive sector entrepre- cies of education have also been important. neurs. However, Tucker et al. (2010, ix) argue Thus, over the years structural changes in that, ‘because conflict [within the party] is the economy have led to new opportuni- largely defined by personality-based struggles ties for wealth creation and accumulation, for senior party leadership and cabinet posts, which in turn change the power structures the likelihood that factional rifts will lead to that are based on such wealth (Khan 2010). a break-up of the CCM appears slim.’ This It is this transformation in power relations is a plausible assessment. The party seems to that contributed to the changes in factional be able to keep together very fluid factions power within the ruling coalitions over time. across regions and social classes. Hence the Furthermore, historically education has also paradox of Tanzanian politics: the party is had a significant impact on elite formation strong enough to keep factions together, but and therefore on the membership of ruling too weak to impose discipline on them.21 coalitions in Tanzania, as has often been the The main reason why such factional con- case elsewhere (Easterly 2001; Banerjee and flicts have not (yet) resulted in a major break- Duflo 2007). up of the party is, according to Tucker et al., that ‘the CCM remains a coherent organi- 3.1 Structural changes in the Tanza- nian economy 20 ‘3 bodies pull out of private lobby group’, Daily News, 1 December 2010. Development entails structural change in a 21 Khan and Gray (2006, Sections B.3.ii and B.3.v) came to a country’s economy. This is a central insight similar conclusion. of the literature on economic development.

18 DIIS WORKING PAPER 2012:06

Structural change entails a diversification high-productivity sectors in Asia, but in away from agriculture and other traditional the opposite direction in Latin America products into industrial activities, modern and Africa’ (ibid., 27). services and so on. In the process, overall The Tanzanian case fits this analysis productivity is expected to rise, resulting well. Since independence, the structure in increasing incomes. Growth-enhancing of Tanzania’s economy has changed more structural change has typically been sub- significantly than some statistics reveal. stantial in high-growth countries, for the According to the World Bank (2007, 8), economy will continue to expand as long agriculture’s share of GDP declined from as labour and other resources are trans- 50 per cent in 1990 to 46 per cent in 2005, ferred from less to more productive activi- and this has long been considered as a ties, even if there is little or no productiv- widely accepted ‘fact.’ However, more re- ity growth within sectors. Globalization, cent data, based on revised information on which elsewhere has promoted technology prices and activities, indicate that in 2001 transfer and contributed to efficiencies in the share of agriculture in GDP was only production, appears not to have fostered around 28 per cent according to the Na- that kind of structural change in Sub-Sa- tional Bureau of Statistics (2007, 8). The haran Africa: ‘Labor has moved in the share of industry, according to these new wrong direction, from more productive to estimates, has also increased more than less productive activities, including, most previously thought, as has that of services, notably, informality… That is indeed what including government activities. However, seems to have typically happened in Latin scores of state-owned and private enter- America and Africa’ (McMillan and Rodrik prises were closed down in the 1990s fol- 2011, 1-3). This conclusion contradicts lowing the structural adjustment reform, much empirical work on the productivity- leaving newly unemployed workers unable enhancing effects of trade liberalization to find new jobs. Probably many moved because such studies tend to ignore the im- into the informal sector, as McMillan pact on those workers who are displaced in and Rodrik describe, or into agriculture, the process. even though productivity in this sector Thus from the 1980s economic liber- has remained largely stagnant for decades alization exposed mediocre enterprises (Skarstein 2010).22 to global competition. As a consequence, At the same time rapid urbanisation took most had to shed labour and/or close place, and presently some 30 per cent of the down as they faced capital constraints population live in urban areas. and could no longer compete. The few is now a city of many millions, and its popu- remaining competitive enterprises could lation is growing some 6-8 per cent per year, not absorb this excess labour. As such, while its economic activities account for more ‘[s]tructural change has been growth reduc- than 60 per cent of national income Mwase ing in Africa…. The bulk of the difference and Ndulu (2008, 430). This Tanzanian ru- between these countries’ productivity per- ral–urban migration has not so much resulted formance and that of Asia is accounted for by differences in the pattern of structural 22 Yu and Nin-Pratt (2011) have a somewhat more optimistic change – with labor moving from low- to view on productivity increases in agriculture in Tanzania.

19 DIIS WORKING PAPER 2012:06

from people in rural areas being pulled by 1971, 37). In any case, there were very few the prospects of rising incomes in the urban African graduates at independence.24 The ed- economy, as it has from them being pushed ucational system was segregated by race.25 by a lack of income-earning opportunities in Education continues to be central for proc- the agriculture sector. esses of elite formation. Official policy has This type of structural change has two always been to expand educational opportu- major implications. Generally speaking, as al- nities for primary and secondary level educa- ready stated in Section 1, Tanzania has not tion (and donor funding has played an impor- been successful at significantly reducing pov- tant role in this lately). On the other hand, erty in recent years, despite impressive eco- access to higher education was restricted until nomic growth since the mid-1990s.23 More- the mid-1980s so as to fit manpower planning over, new opportunities for wealth creation needs. Access quotas and affirmative action and accumulation have slowed down as a re- have been used to try to achieve the latter sult of the apparently declining productivity (Therkildsen 2000a, 84-88; McHenry 1994)). of Tanzanian economy as a whole since the It is also significant that national service, 1980s. This has significant negative conse- established in 1965, also contributes signifi- quences for the emergence of productive and cantly to processes of elite formation and technologically capable entrepreneurs, as will intra-elite relations across ethnic groups.26 be explained in Section 4. Service in the military became compulsory in the late 1960s for both male and female sec- ondary school leavers who wanted to enter 3.2 The importance of education for university. As a result, a pool of people were elite formation formally brought together, some of whom Prior to independence, there were two main later occupied positions in the ruling coali- paths to elite status, both based on educa- tion. The importance of national service in tional attainment. One route went through primary education (and secondary education 24 In 1962, only 147 Africans had a degree among the 2100 for a few). These graduates eventually be- people employed in jobs that normally require a university came teachers or clerks, and some then used degree (Van de Laar 1973, 225). The first university in Tanzania was established in 1962 in buildings constructed to become this as a platform for active participation in the headquarters for TANU, the ruling party (Coulson 1982, party politics or voluntary associations. The 223). Prior to that, Tanzanians went to Makerere University. Only one Tanganyikan received a single term in a European other route went through university into pro- university between the wars, as this was believed to expose fessional positions. Occasionally, but rarely, the native population to subversive ideas. Hence the estab- lishment of Makerere University in Kampala. However, in some proceeded directly from university edu- 1950, Makerere only had 41 Tanganyikan students. There were cation into politics (McGowan and Bolland no more qualified candidates (Iliffe 1979, 445-446). 25 Asians and Europeans made up one percent of the popula- tion, but 50 percent of the government’s educational budget was spent on them (van de Laar 1972b, 108). Only 27 percent 23 There is disagreement about the sources of Tanzania’s of men and 8 percent of women had ever attended school in growth. The World Bank (2007) and the IMF (Nord et al. 1957 (Tordoff 1967, 184). 2009) claim that the economic and structural reforms set in motion since the 1980s have contributed significantly. Oth- 26 In 1972 National Service was put under the control of Min- ers reject this (Research and Analysis Working Group 2009; istry of Defence. It appears to have been abolished for a while Hvidt 2011, 51), arguing that growth has been driven by one- but was revived at the end of the 1990s. ‘Submission to the off events, aid, tourism and favourable but volatile external 95th Session of the Human Rights Committee: March 2009’, factors, as well as by extensification in agriculture, rather than http://www.wri-irg.org/node/6509 accessed 20 December by investment-driven growth. 2010.

20 DIIS WORKING PAPER 2012:06

forging personal relations among university with the particular school that students at- students is substantial according to many in- tended (ibid., 62). It appears that economic terviews with state officials. advantage is converted into educational ad- Today the administrative-political elite is vantage, which in turn leads to positional ad- relatively well educated, and many networks vantages, especially in the public sector and are established on basis of secondary school, the political parties. national service and university ties. Howev- er, access to higher education in Tanzania27 is still limited and skewed. Only 51,000 stu- 3.3 Conclusion dents enrolled for higher education in 2005,28 Major structural changes in Tanzania’s econ- corresponding to only 1 per cent of the age omy since independence, and especially since cohort that was old enough to do so.29 This is the early 1980s, have not led to significant in- very low compared to other EPP countries (5 creases in economic productivity. As a result, per cent for Ghana, 3-4 per cent for Uganda, accumulation based on production has been 1 per cent for Mozambique and 7 per cent for highly limited. The CCM’s attempt to under- Bangladesh).30 mine capitalist entrepreneurs has contributed Although the intake for higher education to this as well. Thus the emergence of pro- has grown strongly in the last ten years31 ductive entrepreneurs with sufficient holding – especially after the World Bank (2002) power to influence policy making and imple- changed its hitherto negative approach to mentation arrangements has only slowly be- higher education – most students in these gun. This theme is further developed in the countries come from the better educated rest of the paper. segments of society. Thus, in Tanzania, al- In contrast, the expansion of the national most all students who completed higher edu- educational system since independence has cation in 1999 came from the richest income produced a growing number of well-educat- group, and 85 per cent of them were male ed Tanzanians. Today, as in the past, formal (Morley, Leach, and Lugg 2009, figure 2). educational qualifications are necessary (but Students from poorer districts were also sig- not sufficient) to gain access to powerful po- nificantly underrepresented. The higher ed- sitions in political and bureaucratic office. For ucation system (including the secondary and Tanzanians who seek employment in formal even primary school systems) is increasingly business organizations, such qualifications are stratified so that access is strongly correlated also increasingly needed.

27 Degree or diploma level and above (Morley, Leach, & Lugg, 2009, 58). 4. CRISIS AND CHANGES IN 28 Up from only 3146 students enrolled in the two public TANZANIA’s RULING COALITION universities in 1990 (Morley et al., 2009, 59). OVER TIME 29 This is measured by the Gross Enrolment Ratio: The number of pupils or students enrolled in higher education ex- pressed as a percentage of the five-year age group following This section focuses on changes in the ruling on from the secondary school leaving age (UNESCO 2009, 255). coalition and its relation to entrepreneurs. It 30 (UNESCO, 2009, 239 and 242-246). explains how three crises have had an im- 31 Tanzania now has eight public and twenty-two private uni- portant impact on power relations within versities and university colleges (Morley et al., 2009, 59). the ruling coalition and its relations to eco-

21 DIIS WORKING PAPER 2012:06

nomic entrepreneurs.32 The three crises are vulnerability of others.’33 In Tanzania, social the mutiny crisis of 1964, the SAP crisis of the groups in the coalition ‘were vested in re- 1980s and the CCM’s legitimacy crisis of the gime continuity by more than rent payoffs; early 1990s. they had predictable access to policy-making and political power and stood more to gain than to lose by standing by … [the party] … 4.1 The roots of CCM dominance [which] provided a credible commitment to and its changes over time these coalition partners.’ In Section 2.1, it was argued that the CCM’s Equally important, the central organs of power is based on three complementary the party have maintained substantial influ- factors: (a) a strong party organisation; ence on who gets to be nominated to run for (b) close links to the bureaucracy and the political office. As Iversen (2006, 12) explains, military; and (c) legitimacy. How this power ‘the incentives for politicians to campaign on emerged and how crises have changed the a broad party platform depend on the abil- party’s power over time is the focus of the ity of parties to control politicians’ re-elec- following discussion. tion chances.’ In the CCM, the nomination of candidates for parliamentary and council 4.1.1 The party organization elections is tightly controlled. The highest or- In response to the economic crises of the gans of the party, the NEC and CC, screen late 1970s, which soon became a full blown and nominate all candidates for presidential SAP crisis (Wuyts 1998), the regime under- and parliamentary elections. This screening took ‘a major effort to revitalize the party, includes vetting by intelligence officers, there- eclipsing TANU with the Revolutionary by linking party and military, as explained be- Party (the CCM) and setting new condi- low. Likewise, the party organs at district level tions for the TANU rank and file to transfer screen and nominate candidates for coun- their memberships to the new organization’ cil elections (but the NEC/CC can, and do (Smith 2005, 442). Then as now the CCM sometimes, overrule these nominations). In is one of the best institutionalised political principle, this represents a very strong cen- parties in Sub-Saharan Africa, together with tralization of political power, for once a can- the Botswana Democratic Party, both of didate has been nominated by the CCM, the which have ruled their respective countries chances of winning a seat in an election are since their independence (Basedau and Stroh very high indeed (see Section 4.1.3). 2008, 23). This institutionalisation is an im- In addition, there has been substantial portant basis for the CCM’s power. Smith continuity in the composition of the party’s (2005, 449), for example, argues ‘that it is the central organs until the recent purge (Sec- origins of, rather than the existence of, sin- tion 2.1). This is partly due to the inclusion gle-party rule that account for the long-term of so-called elders (‘Waasisi/Wazee’: Party viability of some regimes and the surprising Founders/Elders) in the CCM. In addition,

32 Such shifts in relative holding power between factions also 33 He compares parties in four countries (Guinea-Bissau, changes how resources are used to build and maintain the Indonesia, the Philippines, and Tanzania) and finds interesting ruling coalition and to win elections and this issue is dealt parallels in the factors that account for the powerful and long- with in section 5. lasting position of the CCM and of GOLKAR in Indonesia.

22 DIIS WORKING PAPER 2012:06

the party’s grooming and replacement proc- between elders and youth members, which ess is carefully managed. Most of the ‘CCM does not seem to be about ideology but about elites over time are persons that were brought positions and power; (c) a conflict between up through either membership in the parent the party political elite, the party bureaucratic party’s organs or its affiliated organisations,’ elites and the state/party elite to control the especially the CCM Youths League, the CCM party; and (d) organisational conflicts be- Women’s Organization and the CCM Parents tween the Zanzibar and mainland wings of Association (Mmuya 2009, 22-23). Added to the party. this is the fact that important social links be- Mmuya’s observations still largely seem to tween members of the political, bureaucratic hold, although changes have occurred. The and military elites have been built up through SAP crisis of the 1980s started a withdrawal their education in a few high-quality public – albeit uneven – of the state from commer- secondary schools, a single national university cial activities (such as trade and agricultural (until fairly recently) and army conscription of cooperatives) and the privatisation of state- university graduates (Mmuya 1998; Therkild- owned enterprises. This reduced the possi- sen and Tidemand 2007, section 2.1). What bilities for patronage that the party and the emerges is a picture of a fairly close-knit bureaucracy had used in the past. Moreover, group of people at the centre of the CCM, the legitimacy crisis in the early 1990s, which with often close personal relations emerging resulted in the reintroduction of competitive from many years of work in the party. elections in 1995, meant that the party organ- It adds to the relative strength of the CCM isation could no longer depend on the gener- that the holding power of excluded factions ous government subsidies of the past to run is fairly weak, although opposition parties did its organisation. These issues are dealt with in relatively well in the 2010 elections compared Section 5.1. to their poor performance in earlier elections. As Mukandala et al. (2005, 5) explain, ‘over 4.1.2 Close links to bureaucracy, military and 90% of the talented and experienced politi- societal groups cians decided to remain in the CCM under Party- and coalition-building activities to es- multi-partyism.’ Consequently the party’s so- tablish strong ties with important organized cial base cuts across regional, ethnic and so- social groups, and with the state bureaucracy cial divides. The party is therefore a loose co- and military, are central to maintaining the alition of competing factions. This generates power base of the ruling coalition. TANU conflicts within the CCM. Mmuya (1998, 67) embarked on these activities well before in- identified four sources of conflict resulting dependence and continued to strengthen the from: (a) an ideological ‘double vision’ (be- party and its organisation after that (McHenry, tween supporters of a market-led economy 1994). Generally, all significant sources of po- and reform, and those who argue for a con- litical, military, economic and social power in tinuation of -inspired polices, with the Tanzanian state were included or brought an emphasis on state-led development and under party control. As Wangwe (2010, 38) egalitarianism);34 (b) a generational conflict writes: ‘One consequence of the one-party system in Tanzania was to co-opt major play-

34 McHenry (1994) makes the same point. See also Ewald ers in society into organs of the Party. The (2011, Appendix 1). system basically worked on loyalties and pa-

23 DIIS WORKING PAPER 2012:06

tronage relation—CCM (the Party) being the arrived in 1992 – the tilting of the playing parent, using parastatals, cooperatives, mili- field of political competition. These issues tary and the bureaucracy as its instruments. are taken up below. Until the early 1990s secretary generals of cooperatives and trade unions, the chairman 4.1.3 The ability to win elections (albeit on an and managing director of National Bank of ‘uneven playing field’) Commerce, as well as senior members of the Khan (2010) argues that it was donor impo- army or SCOPO (Standing Committee on sition and the threat of a withdrawal of aid Parastatal Organisations) have been all part which forced the ruling coalition to reintro- of the Party’s National Executive Committee duce legislation to allow multiparty competi- (NEC), the centre of political decision-making tion in 1992. He also argues that there was in Tanzania’ (see also Fischer 2006, 577).35 little internal pressure for this change, as Formally, that control ceased following the evidenced by the fact that opposition parties legitimacy crisis of the early 1990s and the have done badly in subsequent elections. change to multi-party competition. Many of This is correct as far as it goes. Donor the ‘new generation organizations’ are offi- funds did decline in the early 1990s (see Sec- cially independent of the party and govern- tion 5). Sharp disagreements had surfaced ment, but also face ‘enormous pressures for over the management of aid programmes, cooptation’, e.g. through the threat of dereg- rampant corruption and poor tax collec- istration or by becoming members of state- tion (Wangwe and Arkadie 2000, 102). But it initiated umbrella organizations (Lindemann was party leaders’ reaction to the perceived and Putzel 2008, 31). Analyses of association- legitimacy crisis of the CCM that was deci- al life in other sectors arrive at similar results. sive. The party could no longer mobilise the The CCM has clearly managed to maintain a population as TANU had been able to do; high degree of control – also after the intro- it had lost its legitimacy. Former President duction of political liberalisation. * Nyerere argued forcefully that, by allowing It is such long-lasting ties to societal groups competition, the party would be able to re- – and the party machinery built up over dec- vive its popular base and therefore be better ades – that has made it possible for the CCM positioned to build socialism. The leadership ‘to react assertively to crises and to prevent also took the decision on pragmatic grounds, economic shocks from pulling [its] coalitions not wanting to resist the wind of change that apart.’ For such ‘ truly institutionalized rul- had swept eastern Europe and the African ing party apparatuses helped greatly to pro- continent (McHenry 1994, 66-67). And many vide a predictable set of political incentives in the CCM were simply convinced that they that signalled to coalition partners that their could maintain their hegemonic position in a interests would continue to be looked after’ multiparty environment (Peter 2002). (Smith 2005, 449). However, this conclusion The latter assessment has, so far, turned omits important additional factors that help out to be correct. Clearly, winning all elec- to explain the CCM’s continuing hegemony: tions at all levels since 1995 with substantial the use of coercion to control opposition to margins helps to boost the CCM’s present its power, and – when political liberalisation power. Thus, the CCM candidate for presi- dency won 62 per cent of the vote in 1995, 72 35 On the relations of the party to CTI, see section 4.4.2. per cent in 2000, 80 per cent in 2005 and 63

24 DIIS WORKING PAPER 2012:06

per cent in 2010. The parliamentary elections Nyerere (1969, 291-294) saw the union be- have also been won by large margins: 80 per tween Zanzibar and Tanganyika in 1964 as an cent of seats in 1995, 87 per cent in 2000, 89 important step towards African unity. It was per cent in 2005 and 78 per cent in 2010.36 also intended to help stabilise the region dur- In local-level elections CCM dominance is ing a tense Cold War period (Southall 2006, even greater. Aggregate information is only 236-238). However, the mainland paid a high available for the 2005 elections, but it shows political price for this. Even though Zanzibar that the CCM won 93 per cent of the seats accounts for less than five per cent of the un- in local governments and 98 per cent of the ion population, it is heavily overrepresented seats in village governments (Tidemand and in the Cabinet, National Assembly and party. Sola 2010, 23; Therkildsen and Geelan 2009). Moreover, the politics of the union be- Opposition parties have virtually no base at tween the mainland and Zanzibar has always the local level. The process of district middle- been troubled, all the more so with the po- class formation – districtization, as Gibbon litical liberalisation of the 1990s. Thus, elec- (2001 842) calls is – ‘has occurred through the tion results in Zanzibar are much closer, and consolidation of business (and often political marred by much greater violence and fre- roles) of (former) district administrative offic- quent irregularities, than those on the main- ers, and through a growing number of retired land (Cameron 2002). The 2000 elections, for or retrenched…administrators…resettling in example, led to riots in Zanzibar in January their home districts.’ The local election results 2001, during which security forces shot into therefore show that at present the party has a crowds of protestors, killing 35 and injur- strong base among such voters. ing 600 according to Human Rights Watch.38 This account of elections is, however, in- Violence erupted again in after the 2005 elec- complete without paying attention to Zanzi- tions, though the 2010 elections were rela- bar. For as Mukandala et al. (2005, 33) argue, tively peaceful.39 ‘the Union [between the mainland and the is- These elections indicate that an opposition lands] continues to feature prominently in any party may eventually win political power on serious discourse on Tanzania.’ The reason is the islands. In the 2010 presidential elections, that the role of Zanzibar is important for the for example, the CCM’s candidate scored CCM. While the ruling party regards the un- 50 per cent of the vote against the opposi- ion with Zanzibar as a central part of both tion candidate’s 49 per cent. The Civil Unit- Nyerere’s and the party’s historical legacy, it ed Front (CUF) also won 22 of the directly is nevertheless divided on how to handle the elected seats in the House of Representatives many politically sensitive problems that this compared to the CCM’s 28.40 Several observ- union generates.37 ers argue that CUF could have won earlier elections had they been free and fair (Liviga

36 ‘Seats’ exclude appointed members. See http://africanelec- tions.tripod.com/tz.html (accessed 30 December 2011). The 38 http://www.hrw.org/en/news/2002/04/09/tanzania-zanzi- turnout for presidential and parliamentary elections as a per- bar-election-massacres-documented (accessed July 2011). centage of the voting age population has increased from 1995 to 2005 (Ewald 2011, Table 9) – although it did drop signifi- 39 http://www.eisa.org.za/EISA/pr20101103.htm (accessed cantly in the 2010 elections according to the website. July 2011). 37 McHenry (1994) gives an excellent account of the Zanzibar 40 http://africanelections.tripod.com/zanzibar.html (accessed issue up to the early 1990s. July 2011).

25 DIIS WORKING PAPER 2012:06

2009b). This has pushed the CCM into pro- limit for the presidency, which reduced presi- longed discussions about power-sharing ar- dential patronage – this has changed. Thus, rangements on Zanzibar. Agreement was fi- 56 per cent of the elected members of parlia- nally reached in November 2009 (Tucker et ment in 1995 also won a seat in 2000. From al. 2010). 2000 to 2005 that percentage fell to 45 per Zanzibar’s relations with the mainland also cent, and from 2005 to 2010 to 43 per cent. remain a sensitive issue. In 1993, for example, The survival rates for politicians on Zanzibar 55 CCM parliamentarians proposed consti- are even lower.43 tutional amendments to establish a separate Although the multiparty elections (except government on the mainland. This was viewed on Zanzibar) have all been declared ‘free as a serious challenge by the CCM leadership and fair’ by some local and international ob- (Mukandala et al. 2005, 18). And in 2008 servers (Liviga 2009b, 19, 22-23), this is an Prime Minister was heavily over-simplification. The playing field is un- criticized by both the CCM and the CUF on even. Liviga argues, for example, that only the islands for declaring that there would be two of the repeals or amendments of forty ‘nothing like the sovereignty of Zanzibar in laws that blurred the separation of powers or the Union Government unless the Constitu- restricted or impinged on political and civil tion is changed in future,’41 thereby seemingly liberties have been implemented. These were questioning Zanzibar’s sovereignty over its recommended by the Nyalali Commission44 internal affairs.42 of 1992 to pave the way for the multiparty This ambiguity in Zanzibar’s status, com- system. Moreover, the amendments to the bined with the real risk of the CCM losing constitution in 2000 were biased in favour power in the islands in future elections, is an of the CCM.45 Since the 2010 elections there important backdrop to understanding the po- have been renewed demands for a revision litical economy of Tanzania. of the constitution, especially concerning the The party’s concerns about winning elec- National Election Commission. tions are shared by individual politicians, for Added to this is the CCM’s use of intimi- life as a politician has become more uncer- dation. A warning in 2007 by the minister in tain over time. During the Nyerere years, as Donge and Liviga (1986) have shown, few 43 Based on an analysis of the names of MPs in the respective were forced out once they had been admitted periods. This does not account for MPs who decided not to run for the second period or who died or withdrew for other to the inner circles of power (some were left reasons. The 2005 to 2010 figure is from Twaweza (2011). out in the cold for a while if they had ‘mis- 44 Set up to analyse the pros and cons of single and multiparty behaved’, but most were re-admitted later). system for Tanzania. With the increase in political competition af- 45 They included (i) election of the President on a simple in- stead of a majority of votes; (ii) Presidential power to appoint ter 1992 – and the introduction of a two-term 10 MPs; (iii) allowing ‘traditional hospitality’ (takrima) in elec- tions making the detection of corruption more difficult; (iv) permitting (potentially costly) house-to-house canvassing in election campaigns; and (v) raising the deposit for election 41 ‘Zanzibar: Premier under fire on Zanzibar status’, The Citi- petitions from Shs. 500 to Shs. 5 million thus raising the fund- zen, 10 July 2008. ing requirements (Mukandala et al., 2005, 3). The CCM domi- 42 A recent dispute about the rights to the continental shelf, nance is maintained by only allowing party members to run where oil and gas may be found, is another example of the for grass-root elections. The CCM’s organisation at this level tense relations between the mainland and the isles. See “East is much stronger and its geographical cover more widespread Africa: Zanzibar MPs protest Tanzania bid for more sea.” East than those of the opposition parties, so that the CCM candi- African, 18 February 2012. dates are often un-opposed.

26 DIIS WORKING PAPER 2012:06

the President’s Office responsible for political For decades the ruling coalition has shown its affairs and civil society relations exemplifies resilience and its capability and willingness to this: ‘opposition leaders are forgetting that use its control of the state against emerging the freedom they are enjoying is not available groups that might threaten its power. The full in other countries…they can say and do what impact of the uneven playing field on elec- they want because the CCM is tolerant and tion results is impossible to assess, but it is can withstand them; …[opposition leaders undoubtedly part of the explanation for the are] inciting the public against their elected CCM’s success. government; …[the CCM] and its govern- ment are the only custodians of this democ- racy’ (quoted in Liviga 2009b, 27). These are 4.2 The Bureaucracy not just empty threats, as interviews with All three major crises since independence opposition politicians clearly indicate. There have contributed to significant changes to are many recorded cases of police arrests the power of the bureaucracy. The mutiny and harassment of opposition leaders during crisis meant that military officers gained ac- campaigns. Police refusal to grant permission cess to important positions in the public sec- for opposition rallies and violent dispersals tor (see Section 4.3). The SAP crisis of the of legal or illegal gatherings are also fairly 1980s resulted in significant changes in the common both during and between election size and mandates of the bureaucracy47 and campaigns (Kelsall, 2003, 61). also helped to set in motion a slow process CCM control of and support from the bu- of land reform which is still unfinished. The reaucracy during election campaigns should legitimacy crisis of the 1990s helped to push not be underestimated either. Although sys- a devolution of powers on to district councils tematic information is lacking, there is cir- and to enhance the role of the bureaucracy in cumstantial evidence to suggest that govern- mobilising voters for the ruling party during ment resources are used in favour of the CCM elections. This is explained below. during election campaigns (Liviga 2009a), and In response to the SAP crisis, central some indication that the party now measures government employment was reduced the performance of key civil servants on how from 355,000 in 1993 to 264,000 in 1998 well they do in mobilising votes for the CCM, ((Therkildsen 2001, 16 & Appendix 3). Em- with the risk of transfers or sackings if they ployment started to grow from then on, do not.46 reaching some 330,000 by 2006, mostly as a Nonetheless the CCM did lose some seats result of donor funding to reach the Millen- to the opposition in the 2010 general elections, nium Development Goals (Therkildsen and and the CCM majority in the presidential vote Tidemand, 2007). The earlier loss of jobs in did shrink, but this is not necessarily an in- government as a result of SAP-induced civil dication of a longer-term weakening of the servant retrenchment and the privatisation ruling coalition. The opposition also did well of parastatals (see below) has subsequently in the 1995 elections, but lost strength in the been compensated for by a rapidly grow- following elections due to internal conflicts ing number of jobs in the social sectors. (Kelsall, 2003, 59), and this can happen again.

47 Resource issues, including pay in the public sector, are ana- 46 Interviews November 2010. lysed in Section 5.

27 DIIS WORKING PAPER 2012:06

The state remains the single most important 56 non-commercial (e.g. institutions of source of work in the formal sector. higher learning).48 Thus, until the end of the 1980s most Following the economic crises of the early graduates found a job in the rapidly expand- 1980s and the subsequent implementation of ing public sector. This was made possible by structural adjustment, access to state-provid- three major economic policy initiatives. Na- ed jobs, especially in the parastatals, shrunk tionalisation of major industrial, agricultural and ‘the elite in toto’ could no longer be re- and commercial properties began with the produced (Kelsall 2002, 610). Those who had in 1967. A Basic Indus- been retrenched or forced into retirement try Strategy (BIS) centred on state-owned may have become active in the productive enterprises (SOEs) was introduced in 1975, sector and in that way helped to build up an aimed to boost industrial growth and em- indigenous productive class of entrepreneurs. ployment and to transform the economy. This is discussed in Section 5.4. And finally, a significant expansion of the The SAP crisis also precipitated land re- cooperative movement was made in order form, although it took many years to get it to cut out the middle man in trade. Subse- off the ground. Villagisation in the 1970s quently, at the end of the 1970s, trade per- had weakened the informal traditional pow- mits were introduced to control the internal er structure. Instead, village governments movement of goods. State trading agencies received substantial powers of allocations in both international and domestic trade over the seventy per cent of land which is were established. classified as village land. In an effort to stim- Through these initiatives numerous job op- ulate production in the early 1980s, the party portunities were created for educated indige- began to encourage large-scale private farms nous Tanzanians as managers of state-owned while at the same time protecting villagers’ firms, trading companies and cooperatives in rights to land through village titling, but cor- the economic sectors. At the same time the ruption undermined implementation. More- emergence of economic entrepreneurs and a over, disputes within villages undermined capitalist class, which was, as is still the case collective control over land, and some be- today, dominated by Asian-Tanzanians and came landless as a consequence. The Land foreigners (see Section 4.4), was stunted. It Tenure Act of 1992 sought to rectify that was the indigenisation agenda that motivated (McHenry, 1994, 92). the state elite to push for these changes, as In further efforts to stimulate production, mentioned earlier. the new 1999 Land Act aimed to speed up As a result the public sector expanded. the commodification of land, to demarcate Thus by 1974 some 80 to 85 per cent of village boundaries, and to issue villages with large and medium sized firms were state- certificates where border disputes with neigh- owned (McHenry, 1994, 131), and the pub- bouring village had been settled. Moreover, lic sector’s share in manufacturing employ- village councils were given the power to ti- ment rose from 16 per cent in 1967 to 53 per cent in 1982 (Skarstein and Wangwe 48 Mozambique was the only country in the region with a 1986, 19-20). By 1988, Tanzania had 413 larger number of state-owned enterprises – over 1,000 – while Senegal had 188, Ethiopia 180, and Ghana 130. Kenya, public enterprises plus some holding com- Uganda and Zambia had even fewer (Temu and Due 2000, panies, of which 339 were commercial and 693).

28 DIIS WORKING PAPER 2012:06

tle the right of occupancy for individual land pears to have become more volatile. For ex- holdings (Kironde 2009; Khan 2010). Despite ample, ministers of agriculture and industry this, land allocation processes continue to on average held the position for less than two cause many conflicts between villagers (Lerise years, while the permanent secretaries in the 2005). Khan (2010, 33) states that generally two ministries lasted around eighteen months ‘there are frequent disputes between villagers on average.50 We do not know if such chang- and investors over land use.’49 Often conflicts es merely reflect the practice of circulating arise because the right of occupancy can be individuals between posts at the same level, withdrawn by the authorities if the land is not or if they actually are moved to less influen- developed (and precisely what that means is tial posts. In any case, such frequent changes open to wide formal and informal interpre- must weaken the substantive sector insights tation). Higher level authorities at district or of office holders in the ministries. national level also become involved where Structural changes within the bureaucracy larger pieces of land are at stake. This often have also been spurred by the legitimacy cri- adds to the complications and to the oppor- sis. One of the most important reforms was tunities for patronage (Sundet 2004). focused on local governments. Starting in the However, in many cases it is the mandate late 1990s and dubbed ‘decentralisation-by- of the village council (typically a few persons devolution’, its formal aim was to move power around the village chairman) to deal with land downwards in government and to make the allocations and conflicts, although the village local authorities accountable to democratically assembly and other organs at the village level elected councils. In practice actual devolution are also supposed to be involved. In practice, of power to the authorities has been limited, such organs are often bypassed and decisions as they depend on central government (or are made by village leaders only. Since almost donors) for most of their revenues. They are all of them are CCM members (see Section not permitted to employ their own executive 4.1.3), the lower level of the party and the either (Tidemand and Sola, 2010). That said, bureaucracy (in practice the lines between the reform has given substantial power over the two are often blurred at this level) has ac- implementation to the local authorities, even quired substantial power in dealing with the in the case of activities that are fully funded crucial asset of land. Khan (2010) argues that by central government. De facto, therefore, the power of the national level of the ruling council politicians and staff have considerable coalition in Tanzania is potentially threatened influence on the day-to-day implementation by this power of the lower level factions over of government activities in major sectors (pri- land allocations. mary and secondary education, water supply, The legitimacy crisis of the early 1990s also health services, agriculture, rural infrastruc- had a significant influence on the bureaucracy. ture, etc.). As the ruling party controls most Life as a minister and as a top civil servant ap- local authorities and almost all village govern- ments, the lower levels of the CCM have sub- stantial powers over this implementation. 49 Surprisingly, and as a warning against generalisations, Cook- sey (2010a, 10) concludes differently. In a case study of horti- Finally, it is important to stress that the de- culture in the densely populated , he found that centralisation reform has also affected minis- ‘… present investors have not experienced major challenges to their land rights from officials, politicians, businessmen or local populations.’ 50 Based on information for the 1995-2010 period.

29 DIIS WORKING PAPER 2012:06

terial mandates. Their role in a decentralised ment and party posts and as managers and system of government is, in principle, to de- directors of parastatals. This ‘militarization sign policies, set standards and monitor the of the bureaucracy’ gained pace from the implementation of government initiatives 1970s so that the Tanzanian administration through local governments. This division of had increasingly come to look like a ‘civilian- labour is, however, fraught with conflicts over military coalition’ by the late 1980s. Finally, resources, as different factions of the central this patronage of army staff has been com- and local bureaucracy fight for control over bined with relatively generous access to state jobs, rents, aid and revenues (Therkildsen resources, as shown in Section 5 (Lindemann, 2000b). 2010; Lindemann and Putzel, 2008).

4.3 The Military 4.4 Emerging entrepreneurs The crises resulting from the failed army mu- A time perspective on three important pro- tiny of 1964 and the war with Idi Amin in ductive sectors – agriculture, industry and 1978-9 jolted Tanzanian army officers – and mining – is provided in the following sec- the army itself – into the ruling coalition. The tion. The discussion focuses on the (possible) failed mutiny was suppressed with the help emergence of indigenous entrepreneurial of British troops. It was preceded by the vio- groups in each of these sectors. It shows that lent overthrow of the government in Zan- historical legacies and the SAP crisis in par- zibar. Both events exposed the vulnerability ticular are important for understanding this. of the party, which responded by attempting For the approximately fifty years after in- both to co-opt and control the army, along dependence an indigenous African entre- with its top officers. Furthermore, the suc- preneurial class has remained weak: limited cessful ousting of Amin cemented the army’s in number, technological capabilities and or- position within the ruling coalition. Not only ganisational strength. That the informal sec- did this give the army influence over party tor accounts for some 60 per cent of Tanza- policies, as has already been described, it also nia’s GDP (World Bank 2005, 34) puts this gave it significant influence over the imple- into perspective. An important explanation mentation of policies, as army officers took for this is that the colonial regime deliberately up positions in the party and the bureaucracy discriminated against African entrepreneurs, (Cooksey 2010b; Lindemann 2010). as has the post-independence regime. Eco- As a consequence military staff have been nomic nationalism, providing jobs for indige- integrated into the party and the bureaucracy nous Tanzanians who did not have the capac- over the years. Prior to the reintroduction of ity to start their own businesses and compete multi-party competition all soldiers were re- with the Asian- and Arab-Tanzanian busi- quired to join the party. A party structure was nessmen, quite as much as socialist aspira- established in the army, and conversely army tions, has driven these changes (Shivji 1975; staff were given seats on the NEC, which Wangwe 2010, 11). The Arusha Declaration enabled them to participate in the formula- should be understood from this perspective. tion of party policies and decisions. Military The impact of SAP has also been signifi- personnel were also increasingly appointed cant, both in terms of weakening the over- to positions in the civilian sphere, to govern- all productivity of the economy (Section

30 DIIS WORKING PAPER 2012:06

2.1) and because it set a slow process of ac- 1980s.’ In addition, economic liberalisation, cumulation in motion. Many politicians and which began in the early 1980s as a result of state officials became gradually engaged in the SAP crisis, aimed to increase productiv- commercial activities. This is signified by the ity through specialisation. Instead it has led Zanzibar Resolution of 1991, which allowed to income diversification and subsistence fall- members of the CCM to engage in what the back due to the risks of relying on the vagar- Arusha Declaration characterised as ‘capitalist ies of a largely unregulated market (Skarstein activities’ such as earning more than one sal- 2010, 99, 103-104). ary, taking up directorships in private firms, The SAP was strongly pushed by donors but having shares in private enterprises, renting also had some support in the ruling coalition out houses, etc. In effect, the declaration for- (Holtom 2007). Nevertheless, it took fifteen malised the accumulation practices that party years to reach a common agreement about and government leaders had already been the content and implementation of many engaged in for many years. According to Pe- reforms (White and Dijkstra 2003) Skarstein ter (2002) the Zanzibar Resolution formally 2010). This drawn-out process shows that killed the 1967 Arusha declaration. In 2009 donors do have significant influence on the President Kikwete tried anew to separate general parameters within which agricultural business from politics. Among other things policies are made, as Hyden (2005) argues, he proposed that politicians with business in- but factions of the ruling coalition still have terests should let these be managed by a trust substantial influence on the implementation while they were in political office (REDET of policies. 2010). That proposal has apparently got stuck Accumulation in agriculture was at first in the party due to internal conflicts. slowed down by the Ujamaa policies, but In the meantime, some accumulation has later, it must be assumed, spurred by the occurred across various sectors. Depending economic liberalisation of the 1980s. Un- on the extent to which these emerging entre- fortunately, there is little information about preneurs are organised or the individuals con- how fast this process is actually occurring. In cerned control sufficient resources, they may 1994, only around one per cent of 3.8 million begin to gain power to influence policy mak- smallholders employed permanent wage la- ing and implementation arrangements. These bour.51 This indicates that landlessness is not trends are tracked below. yet a major countrywide problem, although it is in some parts of the country (Mbilinyi and 4.4.1 Agriculture Nyoni 2000). The legacy of Tanzania’s villagisation pro- Who benefited from the privatisation of gramme of the 1970s and early 1980s has been state-owned farms is also unclear, as dis- significant (Ponte 2002). It has contributed to cussed in Section 5. Today, commercial farms the fragmentation of land holdings so that around urban areas or in areas of high eco- Tanzania is now classified among countries nomic potential are typically owned by ab- with ‘unsustainable’ small farm sizes (United sentee civil servant or politician landlords. Republic of Tanzania 2008, 60). Moreover, There is also limited information about who villagisation ‘was possibly the most impor- tant sole cause of agricultural stagnation in 51 National Sample Census of Agriculture of 1994/95 quoted the period from the mid-1970s to the mid- in Skarstein (2010).

31 DIIS WORKING PAPER 2012:06

actually own larger farms today and what the bargaining, but the extent of authoritarian land is used for, but the number of privately coercion in this should not be underestimat- owned large-scale farms is still low, only 825 ed (cf. Wangwe, 2010, 9), as discussed in Sec- in 2003.52 A substantial indigenous Tanza- tion 4.1. nian commercial farming class is only slowly Political liberalisation in the 1990s opened emerging. Their accumulation may be driven up the de-linking of interest organisations as much by land speculation as by agricultur- from party control, but no strong producer al production (Shivji 2009), although this is organisations have emerged so far. Many poorly researched. NGOs are active in organising small-scale There are many rumours about land ac- farmers, but few if any of them seem to quisitions by foreign investors, but it is dif- have much formal or informal influence on ficult to separate fact from fiction. For exam- national-level decision-making. Large-scale ple, land for producing staple foods such as farmers do not have a specialised organisation maize and rice is sought after by many for- of their own. The Tanganyika Farmers’ Asso- eign governments and investors (Therkildsen ciation, originally established as a division of 2011). Moreover, over four million hectares the Kenya Farmers’ Association in 1935, was of land have been requested for bio-fuel in- registered as a separate entity in 1955 and was vestments, particularly jatropha, sugarcane at that time an interest organisation for set- and oil palm (Sulle and Nelson 2009, 3), al- tler farmers, but it also ran a supply service though only 640,000 ha have been allocated for them. The TFA was nationalised in 1967, so far, and only about 100,000 ha of these and when it was revived in the early 1980s have been granted formal rights of occupan- as a company, it did not formally engage in cy. Proposals for bio-fuel projects involving lobbying. MVIWATA (Mtandao wa vikundi vya initial investments of up to USD 1 billion, wakulima Tanzania) is the largest membership- or several billion USD over the next 10-20 based farmers’ organisation open only to full- years have been made. Maghimbi et al. (2011, time farmers. Started in 1993 with support 49) see this and other examples of land-grabs from the French government, it has expand- as evidence of a process of accumulation by ed to cover 5,200 active smallholder groups dispossession, which continues ‘side by side in 25 regions with about 70,000 households. with the “mainstream” process of accumula- As an interest organisation it is still weak, be- tion’. However, it is too early to assess wheth- ing thinly staffed and poorly financed (Unit- er these investors will seek – and obtain – sig- ed Republic of Tanzania, 2008, 39-45), but it nificant influence on policy-making. does seem to be keeping some distance from As for domestic agricultural producers, the CCM and the government.53 The TCCIA, they are poorly formally organised. Under established in 1988 and heavily dependent on one-party rule all major agricultural organisa- donor funds, has always been dominated by tions were weakened, absorbed or co-opted traders and is not an important lobby organi- into the ruling party. Lindemann and Putzel sation for agriculture (Heilman and Lucas (2008) see this as a result of inclusive elite 1997). The Agricultural Council of Tanzania (ACT) is an NGO established in 1999 (under another name). It regards itself as an apex or- 52 This is a very low number indeed, given that that a farm of over twenty hectares would count as large (United Republic of Tanzania 2005, table 2 (ii), p. 12). 53 Interview via e-mail, December 2011.

32 DIIS WORKING PAPER 2012:06

ganisation but has a limited number of mem- owned by foreigners or Asians with recent ber organisations and receives substantial Tanzanian citizenship. Industry employed donor support. It has very close relations to only 75,000 people, more than half of them government, as its chief executive is a retired in agricultural processing. Around 70-80 per director of policy from the ministry of agri- cent of total industrial production in 1965 culture. Its chairman is a former manager of (the first year of a detailed survey) was pro- a sisal parastatal. Mr Salum Shamte now owns duced by foreign-owned firms (van de Laar a privatised estate and has good relations with 1972a, 85). the political establishment. Nationalisation of major industrial, agricul- Finally the Tanzania National Business tural and commercial properties began with Council (TNBC) is a newly established fo- the Arusha declaration in 1967. Economic rum for government–business dialogue with nationalism and indigenisation, as much as the World Bank as its midwife. It has the socialist aspiration, drove these changes, as al- president as its chairman. Its members are ready mentioned. Indeed, for a long time the high-ranking government officials and rep- party’s economic initiatives aimed to promote resentatives from the larger business associa- the ‘control of the economy by the indig- tions (such as CTI, TCCIA, and TCME). The enous people rather than by expatriates and TNBC did develop a new strategy for agricul- others non-African in origin.’ Political control ture (‘Agriculture First’ – ‘Kilimo Kwanza’ over resources in cooperatives was also im- in Kiswahili) in 2009. The party leadership plemented (McHenry 1994, 107). subsequently embraced it with enthusiasm As already mentioned, a Basic Industry and directed the government to implement it. Strategy based on import-substitution and The architects were high-ranking civil serv- centred on state-owned enterprises was in- ants and prominent indigenous businessmen. troduced in 1975. Through economic liber- It was hatched within a very short time and in alisation and the privatization of parastatals, secrecy (in contrast to the ‘participatory’ ap- the SAP crisis of the early 1980s opened proach that the government officially favours the doors to foreign direct investment and when larger programmes are prepared). This primitive accumulation by local beneficiar- is the only major example in Tanzania of a ies. This process was driven more by donors’ close state–business relationship that has re- threats of economic crises than by pragmatic sulted in some efforts to implement it. How- considerations about future industrial policy, ever, the agriculture ministries (which were and it led to the virtual de-industrialisation not involved in the preparation of the initia- of Tanzania (Arkadie and Dinh 2004). It did tive) are very sceptical so that factional con- probably not help many indigenous Tanza- flicts may derail Kilimo Kwanza, even though nians to become larger entrepreneurs (see it is being spearheaded by the president and Section 5). Most had neither the capital nor the prime minister. the managerial or the technical skills needed to take advantage of this dramatic restruc- 4.4.2 Industry turing. Thus, the 2008 Industrial Survey At independence, there was hardly any indig- shows that there are only 680 manufacturing enous African industrial capitalist group (let companies with more than ten employees alone class) in Tanzania. The few and minor in the country, and that some 60 per cent ‘commanding heights’ of the economy were of manufacturing employment in these is in

33 DIIS WORKING PAPER 2012:06

foreign-owned companies or joint ventures domestic and many foreign-owned compa- (MTIM/NBS 2008). nies as members and has received substan- The SAP-driven economic liberalisation of tial donor support over the years. Moreover, the 1980s, did, however, provide other oppor- all of its chairmen for the last twenty years tunities. It unravelled the earlier attempts by have been prominent indigenous Tanzanian the Nyerere regime to suppress accumulation businessmen-cum-politicians with very good and to promote social equality in the name relations to the ruling party. Even so, several of nation-building. This is well described members interviewed were frustrated by the (McHenry, 1994, Wangwe, 2010). At present, limited policy influence of the CTI. Indeed, primitive accumulation centred in and around while the CTI is active in issues related to the state is mainly related to the possession taxation and wages, it does not seem to have and dispossession of land and mining rights, had much impact on, or interest in, other in- and to real estate deals and trade (Shivji 2009; dustrial policy issues. Kelsall 2002). No systematic study on this has been done, but the impression gained from 4.4.3 Gold mining casual observation is that the main effect ‘has Over the past fifteen years, large-scale, capital- been a resurgence of the Asian business com- intensive, foreign-owned and -operated gold munity, not only in areas where they tradi- mining has been given priority over labour- tionally played a dominant role, such as trade, intensive and local artisanal activities, thereby but also in new areas such as banking and fi- curbing the potential for the development nance, and real estate development’ (Wangwe of a prosperous Tanzanian mining industry and Arkadie, 2000, 67-68). The present forms run by domestic investors and entrepreneurs of accumulation therefore do not seem to be (Cooksey and Kelsall 2011; Bourgouin 2011). precursors of productive capacity, but may ac- This has created a situation in which local ar- tually undermine it as the economic rewards tisanal-scale interests, such as access to land, in other sectors (including in politics) seem so are increasingly being ignored, even though much better. ‘A genuine middle-class ground- artisans continue to be given credence in pol- ed in production and involved in the chain icy and political discourse. of accumulation is constantly stifled’ (Shivji This represents a significant shift in policy 2009, 50). A similar assessment is made by compared to the Nyerere years. He put the Khan and Gray (2006, 55). The increasing ev- limited large-scale mining activities in the idence of political competition and corrup- hands of the state (Chachage 1995). The first tion reflects ‘the growing opportunities avail- comprehensive post-independence mining able to political representatives in managing law was then promulgated in 1979. It vested the non-market transfers and allocations of the State Mining Corporation (STAMICO) assets like land and natural resources.’ with the legal powers to engage in the explo- With the wealth basis for the holding pow- ration, prospecting, research and exploitation er of most industrial firms being weak, it is of mineral assets. It also provided the basis no surprise that the organisational strength for artisanal-scale mining (ASM) in the coun- of industry companies is also limited. The try and gave powers to the Minster of Miner- TCCIA, which also seeks to represent small- als to set aside areas for the exclusive use of scale enterprises, is weak, as already noted. Tanzanian miners. In 1980, the government The CTI, established in 1991, has many large formally recognised ASM as an important

34 DIIS WORKING PAPER 2012:06

economic activity, began to earmark land for reforms. Thus, a sharp ideological shift in ASM operations and introduced trade regula- Tanzania’s policy towards LSM has occurred tions to encourage foreign suppliers of min- between the Nyerere years of nation-building ing equipment to do business with artisanal and state control of mineral assets and the miners (Cooksey 2011). A decade later, the introduction of liberalisation reforms. Most government agreed to start purchasing miner- importantly, the new policy gave room to als from artisanal miners through the Bank of foreign investors in the mining sector, which Tanzania, the National Bank of Commerce had the effect not only of revitalizing the and the Cooperative and Rural Development country’s LSM sector, but also of bringing Bank. in the bulk of FDI into the country in gen- Until the early 1990s, the local ASM sector eral. In 1999, for example, FDI inflows for was therefore given credence in both policy gold mining represented nearly 75 per cent of discourse and political practice. The political total FDI investment in Tanzania. Three in- elite played an active role in addressing the ternational mining companies54 now operate specific needs of the sector. Moreover, large- major gold mines and attract great attention scale mining (LSM) activities remained lim- with their large investments, The significance ited. When STAMICO was set up in Tanza- of the artisanal mining sector, which employs nia in 1972 it inherited a formal gold industry over 500,000 people (ILO 1999, 5), has been that had declined to insignificance, and there overshadowed by this development. Over the was limited active large-scale mining activ- years there has been virtually no local accu- ity at this time. Foreign-owned companies mulation in the gold minerals industry. Why were not allowed to hold mining concessions, is this? which formally impeded the growth of the Large-scale foreign-owned mining com- LSM sector through imports of capital and panies were prioritised in political practice technological expertise. This also meant that by the ruling coalition due to their perceived Tanzania was not able to benefit from the potential in the development and rapid eco- global stimulus to gold exploration and min- nomic growth of the country. It was assumed ing resulting from the liberalisation of the (or naively hoped) that a liberalized mining gold price in 1975. However, in view of the sector could potentially contribute up to 10% country’s considerable mining potential and of GDP and would generate enough wealth poorly performing economy, LSM activities so that the country would no longer be de- were increasingly considered a potentially im- pendent upon, or need, donor support.55 The portant source of tax revenues and foreign privatization of the state-owned mining sec- exchange. tor prioritised foreign companies. Sometimes The country was opened up to exploration this required artisanal miners to be cleared and mining after the adoption of a structural from the land, thus reducing the land avail- adjustment programme in 1986. The legal, able to local artisanal miners. Currently all ca- regulatory and fiscal frameworks were later dastral plots of land available for mining or restructured substantially to create a sector which would be conducive to foreign pri- 54 vate investment, and the Mineral Policy Act AngloGold Ashanti Ltd, Barrick Gold and Placer Dome. 55 Whilst nearly half of Tanzania’s exports are now attributed of 1997 and the subsequent Mining Act of to the mining industry, it only accounts for 3.2% of GDP and 1998 reflected the direction charted by these 3.6% of the Government’s total tax revenues.

35 DIIS WORKING PAPER 2012:06

prospecting licenses have been issued, mak- this area as long as they sold their gold to the ing it impossible for artisanal miners to ex- Government. pand their operations through the acquisition A Bulyanhulu Miners’ Committee was of new licences. set up and applied for prospecting rights All this indicates that the Tanzanian gov- at the site, but its application was never ernment did succeed in gaining the confi- considered. Instead, in 1994, the Ministry dence of the foreign mining companies. In of Energy and Minerals decided to grant contrast, the ASM sector was ignored, de- mining rights to large-scale transnational spite the declared intentions in the 1997/8 corporations, and Sutton Resources Ltd, a and 2010 Mineral Policies and Acts to devel- Canadian company, signed a Mine Develop- op it. Opportunities to support local capital ment Agreement to be operated through a accumulation in ASM were simply disregard- local subsidiary called the Kahama Mining ed. Instead, prospecting licences were issued Co. Ltd. In other words, despite President’s to a plethora of foreign exploration compa- Mwinyi’s decision to support local artisanal nies, and mining licences were given to the mining in the area, the minister overruled three major companies mentioned earlier. his wishes and granted the licence to the But no provisions were made in the reform foreign company. KMCL addressed the process to build backward, forward or lateral courts in 1995 to have the artisanal miners linkages to the local economy through, for remaining in its licence area evicted, but the example, local procurement (this is often an courts did not rule in its favour. important development objective in other Despite this, the artisanal miners were not countries). legally protected. In 1996, then President Clearly, there was a ruling faction which fa- Mkapa showed that his government was com- voured the development of an FDI-led gold- mitted to the development of a large-scale mining industry. However, evidence suggests commercial mining sector when the minister that the ruling coalition was not united on responsible for mining told the artisanal min- this. The events which took place at the Buly- ers to leave, as they were trespassing. Evic- anhulu mine in Kahama District, now owned tions were ordered in July 1996, and the site and operated by African Barrick Gold, are a became the symbol for a new approach to prime example of conflicts within the coali- mining policy for the country, where the large tion over the country’s approach to the devel- foreign-owned companies predominated over opment of gold mining. local artisanal interests. Artisanal miners here began activities in This conflict between FDI-led and arti- the region in 1975, and over the years the sanal mining, combined with a lack of vis- site has attracted them in their thousands ible benefits to socio-economic development (Cooksey and Kelsall 2011; Chachage 1995). from mining privatization, has fuelled not However, when a prospecting licence was only widespread discontent and popular re- granted to the Canadian company Placer sistance, but further disagreement within the Dome in 1989, conflicts began between the ruling coalition over how to direct the mining mining giant and the thousands of local min- industry. Consequently, in 2007 the Govern- ers operating on this land. President Mwinyi ment established a commission headed by visited the area in 1992 and reassured local Justice Mark Bomani (commonly referred to artisanal miners of their rights to operate in as the Bomani Commission) to investigate

36 DIIS WORKING PAPER 2012:06

the mining sector and to review the existing presented above. Both colonial rulers and the mining development agreements with major one-party state suppressed the emergence multinational companies. It did not address of an indigenous entrepreneurial group, the problems of artisanal miners. Instead, which as a result remains weak today. Even the 2008 report made numerous recommen- at present, the ruling coalition’s commitment dations as to how Tanzania could enhance to supporting it remains ambiguous. This is income from the minerals industry such as also reflected in state relations with business increasing royalties and granting fewer tax associations. There are many such associa- exemptions to new foreign investors. Also tions in the productive sectors, but the ma- amendments to effective tax and duty con- jor ones are not particularly strong and tend cessions were proposed. The corollary of this to be timid in their relations with the state, focus on FDI in mining, which is at the root and often dependent on donor financing. It of ideological and economic debates and rifts is difficult to find clear cases of instances within the present ruling coalition, is a con- where their influence has had an impact on tinued neglect of artisanal miners (and in the government policy.56 The small size of the case of Bulyanhulu, a tragic one). formally organised business community and This marginalisation of small-scale min- its diversity provide a further explanation for ers is reflected in the business organisation this lack of influence (Lofchie and Callaghy for the industry. The FDI-dominated indus- 1995). The ambiguous relations between the try is organized through individualized bilat- state and private-sector entrepreneurs pose a eral relations between the government and dilemma for many business organisations in the mining companies themselves. While a steering between cooptation or confronta- Chamber of Mines and Energy was estab- tion in their interactions with the state (Heil- lished in 1994, this structure remains very man and Lucas 1997). weak and only provides artisans with a sym- The analyses have also shown that crises bolic role in the industry. Its members from are important in explaining changes in pow- the gold-mining industry are primarily the er relations between factions of the ruling foreign executives or Tanzanian upper man- coalition and between the coalition and the agement staff of the three major mining emerging entrepreneurial class. These crises, companies and some of the foreign explo- in particular the CCM’s legitimacy crisis in ration companies. Clearly these global gold- the early 1990s, have generally increased the mining giants hold such economic power power of lower level factions of the coali- that they do not need to be organized col- tions, especially during the last ten to fifteen lectively. The problem is that the plight and years. Zanzibar, and its role in the union, re- economic interests of artisans are not being mains an important source of political ten- heard or formally addressed by this chamber sion, with the potential to break up the un- (nor by government). ion with mainland Tanzania.

4.5 Conclusions The importance of historical legacies in shap- 56 See Gibbon (2005) for a selective analysis. The advocacy activities of most associations seem to concentrate on fac- ing the main features of the ruling coalition tors that can reduce production costs and taxes, not on more in Tanzania clearly stands out in the analyses general policy issues for particular industries.

37 DIIS WORKING PAPER 2012:06

Figure 1. Tax collection by TRA department, 2001-2007 (percentage of GDP) ��

�� ������������������ ��� �� ���

��

�

�

�

�

� ������� ������� ������� ������� ������� ������� �������

DRD: domestic revenue department; LTD: large taxpayer department Source: Nord et al. (2009, figure 13).

5. RESOURCE USE TO STAY IN substantial. In shilling terms total revenues POWER grew five-fold from 1996 to 2006 (Tanzania Revenue Authority figures from July 2010). The analyses in this section are based upon The potential ability of the ruling coalition the proposition – outlined in Section 1 – to ‘do things’ has therefore improved signifi- that a ruling coalition will seek to use the cantly since 1995. resources it controls to strengthen its coali- Equally interestingly, Tanzania’s largest tion and to win elections (see also Mesquita companies are major contributors of tax et al. 2005; Kjaer and Therkildsen 2012). revenues (identified as LTD in Figure 1). These resources consist of tax revenues, aid The Large Taxpayer Department handled and rents. the taxes of 410 companies in 2010.58 Such firms paid almost half of total taxes in 2007, up from one-fifth in 2001.59 If power follows 5.1 Taxes and aid money, this indicates that large companies Tax collection has increased significantly in should have substantial influence on policy- recent years, from 10 per cent of GDP in the making and implementation. Yet, the analyses early 2000s to 14 per cent in 2007 (Figure 1) presented in Sections 3.4 and 4.4 show that and 16.5 per cent in 2008 (IMF 2009, 18).57 this is generally not the case. This only further Since annual GDP growth rates of four to six supports the argument that the relationships per cent have been achieved from the mid- of such companies to the state are particular- 1990s (Utz 2008, table 1.3), the growth in the istic rather than formally organised. volume of tax-generated revenues has been

58 TRA list of large firms as of July 2010. 57 Non-tax revenues add another 0.5 percent to the available 59 Small and informal firms contribute little to total revenues revenues. (Hoogeveen and Ruhinduka, 2009, 62).

38 DIIS WORKING PAPER 2012:06

Figure 2. Net official aid to Tanzania and developing countries, 1960-2006 ��� ����������������������������������� ���

���

��� Developing Countries Total �� Tanzania

� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ���� ����

Sources: OECD DAC database; IMF staff estimates.

As shown in Figure 2, aid has continued building and win elections, as illustrated in to be an important financial resource in the following.61 past decades, with volumes of aid growing rapidly in recent years. Aid volumes have risen from some USD50m/year in 1970 5.2 The Party to between approximately USD1000m Assessing the political economy implications and USD1400m per year during the last of how and by whom the party is funded is decade (van Donge and White, 1999 and both important and difficult, for much of OECD data). However, the data in Figure this is done informally or in non-transpar- 2 also show that donors reduced aid flows ent ways. Here the focus is on the financing to Tanzania during the early phase of the of the party itself, and on its use of public SAP crisis (1980-1985), as well as during revenues to build and strengthen the ruling the legitimacy crisis of the early 1990s. coalition and to win elections. Subsequently, aid increased again. In the During one-party rule, TANU (and later 2000s, aid has made up some 10-12 per the CCM) received subsidies from the gov- cent of GDP and a remarkable forty per ernment, approved by the party-controlled cent of the government budget (IMF 2009, parliament. As a result, party funding was figure 2). Moreover, Tanzania receives for- fairly generous. Today, public funding of po- ty per cent of this aid as budget support60 litical parties is regulated by the Political Par- (OECD 2007). As a group, OECD donors ties Act. It states that any party winning ‘at are clearly very important financiers of least five per cent of the total votes cast in Tanzania’s ruling coalition, and they help it, directly or indirectly, to fund coalition- 61 The aid figures underestimate the importance of aid, as there are still some 1600 donor-funded projects in the country, many of which are funded off-budget (Frot and 60 More than any other EPP country. Santiso 2010).

39 DIIS WORKING PAPER 2012:06

the presidential elections and … in addition ciety’ (Tucker, Mukandala et al., 2010, 3) in … a seat in parliament’ qualifies for govern- which Asian business people especially (but ment subsidy. Since the CCM wins elections not only) exchange the funding of CCM ac- by a large margin (Section 3.1), this favours tivities and its election campaigns for rents to the ruling party. After the 2000 elections, for individual companies and individuals. Such example, the party drew ‘a monthly subven- rents often consist of non-market access to tion of US$547,000, while the combined par- property and other assets, as already men- liamentary opposition parties draw less than tioned, but also to licenses, duty exemptions, US$50,000 a month.’ (Sokomani 2005, 87). tax reductions, etc. (Wangwe, 2010). This has This massive advantage in party funding is gone on for years, although interactions have also stressed by Kelsall (2003, 61).62 probably been strengthened and accelerated In addition, with membership up to 3.5 following the SAP and legitimacy crises in the million, membership fees help to finance the 1980s and 1990s. party. The CCM also used to run commercial These views are supported by Cooksey and enterprises (through the SUKITA company), Kelsall (2011, 89) who conclude that ‘[g]rand but has since closed down. By 1998 it had ac- corruption is perpetrated by ad hoc alliances cumulated debts of TZS 12 billion (Fischer of politicians, officials (including members of 2006, 358). The CCM also had its own oil the armed forces), and private brokers, usu- company at the time, which, due to lucra- ally with the tacit knowledge but not direct tive licence allocations and smuggling, was a involvement of the President. Driving these highly lucrative business (Fisher 2006, 382). scams is a combination of individual greed However, Ewald (2011, Appendix 1, 346- and the need to secure funds for ruling party 347) argues that the party still owns a number election campaigns.’ of companies, as well as real estate (office The most spectacular known example of buildings, storehouses and factor buildings) such corruption is the plundering of the Ex- in prime locations in all major cities. The par- ternal Payments Account (EPA) at the Bank ty also owns all major sports arenas (except of Tanzania (BoT), which, it is widely be- the new national stadium) and the Kiswahili lieved, was used to finance the CCM’s 2005 newspaper, Uhuru. Business links with com- election campaign. Ernst and Young, the au- panies owned by the military may also exist dit company, estimated that most of the TZS (see Section 5.4). 130 billion paid to recipients of EPA went Private contributions to parties are not to ‘bogus companies set up for the purpose regulated by law, nor are there any spending of emptying the EPA, an old commercial limits on them. Close but informal relations debt facility transferred from the Ministry of between business people and politicians are a Finance to the BoT.’ Payment of more than prominent feature of patron-client relations TZS 90 billion was ‘based on “invalid and in Tanzania. This is ‘creating a political and fraudulent supporting documents.” Moreo- economic oligarchy at the top levels of so- ver, due to “missing information,” the audi- tors could not establish [the] validity of the

62 In 2008 the government proposed to introduce a Constit- remaining TZS 40 billion’ (Cooksey and Kel- uency Development Fund to allow members of parliament to sall 2011, 27). control funds for community-oriented work. Given the domi- nance of the CCM, this gives its MPs an added advantage. The In the case of Zanzibar, rents have played a proposal was hotly debated and was enacted in 2009. particularly important role in maintaining the

40 DIIS WORKING PAPER 2012:06

CCM’s political power on the islands. Here, as cause voters lack a viable alternative: given discussed in Section 4.1.3, the party’s position the CCM’s continued hegemony, voters can- has become increasingly precarious since the not rely on the opposition for government political liberalisation of the early 1990s. To resources to improve their well-being. diffuse social and political unrest and to help Mesquita and Smith (2009) have a differ- the party continue to win elections, the CCM- ent view on this. They focus on the distribu- controlled Zanzibar government lowered the tion of private goods (e.g. fertilizer vouchers rates on imported goods such as rice, sugar, for maize) and public goods (e.g. health-care wheat, flour, cooking oil, kangas and soap facilities, school provision) across districts (Fischer 2006, 341). This has also turned into based on the presidential vote in the 2005 a lucrative smuggling business: goods are im- elections. They argue that the ‘total vote share ported through Zanzibar to be sold on the for the winning party generally has an insig- mainland, sometimes not even entering Zan- nificant effect on public and private goods al- zibar, as only the documents make the jour- locations’ at district level. However, districts ney for certification purposes. Therkildsen loyal to the CCM63 tend to receive more pri- (2011) argues, in the case of rice, that some vate goods than districts where the opposi- of the businesspeople involved in this traffic tion does well. Here more public goods are help finance the party, which explains why the provided. ‘Zanzibar route’ – one of the most important These analyses are highly problematic.64 The channels of tax evasion and a main reason interpretation of statistical results depends on why it has been difficult for Tanzania to bal- the data used, but the authors do not clearly ance its budget in the past (Fischer 2006, 235) specify their sources, nor the exact types of – has proved so difficult to close down. Highly transfers that they are analysing. Thus, alloca- subsidised electricity rates for Zanzibar serve tions of funds for development expenditures the same political purpose (ibid., 219). for local government (Local Government Through its control of parliament and Development Grants) are formula-based and close relations to the bureaucracy, the party are, in practice, largely followed. Allocations can also use public revenues and aid to fur- of funds for recurrent expenditures (salaries ther its cause. For example, according to and other expenses) are not formula-based, Weinstein (2011), research on competitive but are calculated on the basis of actual staff- democracies shows that the most effective ing, which varies widely across districts: there distributive strategy of a ruling party to win are, for example, far fewer students per teach- votes is to target resources for districts where er in urban areas than in rural ones.65 This party support is weak. Tanzania is, however, skewed distribution is not the result of a de- different. From 1999 to 2007 the CCM tar- geted resources towards districts that voted for the party with the highest margin of vic- 63 Measured as the number of votes for the second largest tory. This is not a result of patronage politics, party (also called the minimum winning coalition). 64 This section builds on discussions with Per Tidemand, for she argues. Rather, the ‘CCM attempted to which we are grateful. send a signal to non-supportive districts that 65 In the early 2000s a larger analysis of local government lower vote shares are punished with lower per staffing was carried out aimed to achieve a more equal dis- tribution of staff across districts. The attempt was abandoned capita expenditures and smaller changes in as teachers and health staff resisted transfers to marginal or budget rates.’ This is a rational strategy be- rural districts.

41 DIIS WORKING PAPER 2012:06

liberate policy but a legacy of decades of al- tion has immediate, visible results, and imple- lowing teachers to work in urban areas, where mentation can be clearly identified with the they prefer to live. It is therefore not the poli- party in power. These characteristics fit a‘logic tics of winning elections which explains the of winning office’ and indicate that political fact that urban areas – where the opposition elites do respond to popular demands for bet- does relatively better – receive higher recur- ter government performance in the context rent grants than rural areas. The third type of of multiparty competition. This is contrary fiscal transfer involves the funding of larger- to arguments in much of the literature on scale infrastructure (bridges, airports, roads, policy making in Africa, which emphasizes irrigation, etc.) or special activities.66 These the importance of factors such as clientelism, allocations are not part of the intergovern- ethnicity, organized interest group influence mental transfer system and may be vulnerable and donor pressure in explaining major deci- to political manipulation, but we do not know sions.69 if such transfers are included in the authors’ It adds to our understanding of the emer- analyses.67 gence of populist election-focused politics in Both papers are also based on another, Tanzania to look at another landmark deci- equally questionable premise. It is assumed sion: inviting foreign direct investment in the that the ruling coalition has sufficiently de- mining sector in 1998 after decades of keep- tailed knowledge to be able to fine-tune allo- ing such investments out. When that deci- cations to support or punish districts accord- sion was taken, it signified an attempt by the ing to district-level voting patterns. That is Mkapa regime to start to exploit the coun- unlikely. The Weinstein and Mesquita statisti- try’s mineral resources in a more intensive cal analyses are therefore highly questionable way. However, the political elite soon realised (see also Section 5.3). that this was seen by many as giving away the The approach taken by Kjaer and Therkild- country’s wealth to foreigners. The debate sen (2012) does not make such assumptions. reflected a strong and widespread sense of In an analysis of landmark decisions68 in Tan- economic nationalism. In response politi- zania and Uganda since 1995, they show that cians now play the populist card in trying to these share certain key characteristics: citizens renegotiate royalties and tax agreements with countrywide are targeted, policy implementa- the mining companies (see Section 4.3.1). In other words, the mining policy was a ‘devel- 66 Such as fertiliser vouchers. These are only allocated to a opmental policy’ that got bogged down in limited number of important maize- and rice-growing dis- populist resistance. tricts, not to all districts, as Mesquita and Smith imply. Here the skewed distribution of vouchers is more about capture by the CCM’s leaders at village and ward level than it is a result of a deliberate election-winning strategy by the CCM (Kjaer and Therkildsen 2012). 5.3 The bureaucracy 67 For the 2010/11 financial year, districts received around The critical importance of the bureaucracy 500 million USD as development grants, and around one in the ruling coalition is clearly reflected in billion USD in recurrent grants. Funding of larger projects is transferred through ministries and government agencies. the budgets and in rent benefits. Civil serv- Their budget information does not show allocations to indi- vidual districts (Per Tidemand, personal communication). 68 Among the landmark decisions are the introduction of free 69 Government over-spending in election years – a familiar primary education, the abolition of local government taxes pattern across the world – is also used in Tanzania, as experi- and fertiliser subsidies. ences from the 2010 elections show.

42 DIIS WORKING PAPER 2012:06

ants in Africa are, generally, the best paid in Kikwete also benefited from the sale. After the world relative to the people they are sup- he became president in 2005, he announced posed to serve, and Tanzania is no exception that certain houses would be renationalised. in this regard (Schiavo-Campo, Tommaso, This has not happened. (Cooksey and Kel- and Mukherjee 1997, table 2). The difference sall 2011, 29). can be seen as a rent that reflects its power. With the decentralisation to local govern- Indeed, the median salary level for public em- ments from the late 1990s, the lower level fac- ployees in Tanzania doubled from 1992 to tions of the ruling coalition received rapidly 2001, while it declined in all other countries increasing transfers of funds from the central included in the study by Kiragu et al. (2004, government. Recurrent expenditures by lo- figure 4.1). Moreover, against the advice of cal governments rose almost three-fold, from pay-reform specialists in the mid-1990s, pri- TZS 1.3 billion in 2001 to 3.1 billion in 2006. mary-school teachers and health staff in par- Their share of total government-recurrent ticular benefited from salary increases. This expenditures rose from 18 per cent to 24 per cabinet decision was clearly aimed to improve cent during that period. In addition, the local relations with these front-line staff, who authorities now spend 17 per cent of the total can help to mobilise voters during elections government development budget (Tidemand (Therkildsen 2000b, 67). and Sola, 2010, 21). Subsequently, pay reform in Tanzania The extent to which such transfers are has lost momentum. Instead, the authori- used to build and maintain district-level fac- ties have turned to allowances to implement tions of the ruling coalition is the basis for their own organization-specific ‘pay reform.’ much speculation. The two analyses present- This is clearly documented by Valentine ed in Section 5.2 do not provide reliable evi- (2006). Today, various allowances are typi- dence of systematic use of public funds for cally equally or more important for incomes patronage and/or coalition-building. This is than salaries are (Therkildsen and Tidemand, not to say that central and local budgets are 2007). Some $390m was spent on twelve dif- consistently used for their formally intended ferent allowances in 2008, equivalent to the purposes – they are clearly are not, as visits salaries of two-thirds of all primary-school to the constituencies of ministers and pow- teachers, and making up 59 per cent of the erful members of parliament easily confirm. total wage bill in 2009. Allowances increased Expenditure-tracking surveys also show that threefold between 2001 and 2006. They typ- there are significant problems with the flow ically accrue to higher level officers and are of resources to primary education, for exam- also clearly used for purposes of patronage ple, and that rural councils are the most dis- (Twaweza 2009). advantaged. The reasons are many (complex Another example of rent benefits to the system, poor management and supervision) upper levels of the bureaucracy is the sale according to Twaweza (2010), a critical NGO. of government houses to senior civil serv- Rent-seeking at the district council and school ants, ministers and high-level CCM offi- levels undoubtedly plays a role toos although cials in 2005. Although many houses were Twaweza does not say as much. But the leaks in a poor condition, they were located in of funds illustrate the factional nature of the prime locations in Dar es Salaam and could ruling coalition. Even in primary education, be sold for a very handsome profit. Jakaya the improvement of which is an important

43 DIIS WORKING PAPER 2012:06

element in the CCM’s strategy for winning za (2009), allowances to the military (national elections, the party cannot control or prevent service) was the fastest growing item across the diversion of funds as it channels its way all sectors between 2007 and 2009. through the bureaucracy and is distributed at Information about the military’s involve- the lower levels. ment in business is very limited. However, Rent-seeking is also prevalent in the mining since the SAP crisis of the 1980s the army sector, as the licensing process offers a unique /JKT has maintained its involvement in eco- opportunity for rents. While there is no single nomic activities. It has, at various points in person to bribe in order to secure a prospect- time, run a number of commercial enter- ing or mining license in Tanzania, there is a prises, such as the Meremeta gold mine, the sizable and inefficient bureaucracy with leak- SUMA JKT tractor assembly factory in Kiba- ages and minor corruption at various points ha, a 4000-hectare rice farm in Kilembero during the process. The waiting time for an and some trading companies. exploration licence is about two years (while The SUMA JKT factory, in collaboration in Australia it is about two hours), and inter- with Jeetu Patel, an Asian-Tanzanian busi- views confirm that these are often handed nessman, has been licensed to import power out to members of the ruling party for pur- tillers, a central element in the government’s poses of speculation. Yet despite their size ‘Agriculture First’ initiative (Therkildsen and financial resources, mining companies 2011). Meremata received more than TZS have not in any sense ‘captured’ the Tanza- 10bn from the External Payment Account nian state.70 There does not appear to be any scheme, which is now at the centre of a major effective central coordination of rent-seeking corruption scandal (Section 5.2). The compa- in the mining sector or more generally. ny was taken over by Deep Green Finance in 2005. However, the CCM Secretary General, Yusuf Makamba, ‘denied that “CCM has any 5.4 The military business links to Deep Green.”’ (Cooksey The importance of the military in the ruling and Kelsall, 2011, 27). coalition is also reflected in government ex- penditures. Military budgets ‘appear to rep- resent a systematic strategy of cooptation’, as 5.5 Emerging entrepreneurs the percentage of the national budget allocat- Productive sector contributions to govern- ed to the military increased from 4 per cent to ment revenues, the use of government funds 14 per cent between 1970 and 1990 accord- in the productive sectors and the various rent ing to Lindemann and Putzel (2008, 29). This transfers between state and business are dealt percentage has dropped since then and now with below. The aim here is to assess the po- appears to be around eight percent according litical economy aspects of these resources. to Cooksey (2010b, Chart 2.1), although the From a sector perspective, agricultural budget is not transparent regarding defence producers contribute only modestly to cen- expenditure. Moreover, according to Twawe- tral government revenues because the great majority of producers are informally operat- 70 State capture occurs if a relatively small number of firms ing smallholders. District authorities do col- are able to extract concentrated rents or to purchase in- dividualized provisions of under-provided public goods from lect various crop cesses, and these can have public officials at various levels. considerable negative impact on production

44 DIIS WORKING PAPER 2012:06

(Cooksey, Shao, and Mkende 2007), but are agriculture. The government increased allo- generally small in aggregate terms. The in- cations to agriculture from USD62.4 million dustrial sector is claimed to contribute half in 2003/04 to USD158.5 million in 2006/07, of the government’s own revenues generated with about 80 per cent of the sector devel- through sales, corporate and income taxes im- opment budget financed by donors (Wolter posed on industries (Mbelle et al. 2010, 30).71 2008, 17). The mining sector contributes only four per The cost of fertiliser subsidies is a very fast- cent of government tax revenues, as some of growing budget item, rising from 4 per cent the larger mines have just turned profitable, of the total agricultural budget in 2003 to 28 and companies utilise the tax deductions that per cent in 2008 and representing a tripling their contracts with the government allow. of tonnage, though the productivity-enhanc- Dissatisfaction with this is widespread among ing effects are unknown (Therkildsen 2011). the public and the politicians, who clearly Allocations to promote small- and large-scale expected foreign direct investment in large- industries are very modest, as is the funding scale mining to produce significant revenues of initiatives to promote manufacturing ex- fast. During the legitimacy crisis in the early ports, for example. Government funding of 1990s, for example, when donors were reduc- the mining sector, including small-scale min- ing aid due to government mismanagement ers, is miniscule. of public funds, several top civil servants saw Tax exemptions are a major type of for- large-scale mining as a fast-track escape from malised rent. Many have been politically con- donor dependency.72 troversial, and several ministers have, over The most interesting fact about productive the years, been forced to resign amidst accu- sector revenue contribution is, however, that sations of favouring specific businesses. All only some few hundreds of large firms con- such scandals seem to have involved imports tribute a large proportion of total revenues. of commodities related to agriculture, indicat- Because these firms operate across many ing that importers of such commodities have sectors and are poorly organised as a group good connections to the ruling coalition (on (formally as well as informally), they do not rice imports, see Therkildsen 2011). engage in collective action in dealing with the In recent years, tax exemptions seem to state. Instead many of them seem to prefer have been on the increase. In 2009/10 alone, to deal with the government on an individual 2.3 per cent of GDP or TZS 695 billion was basis as shown below. granted in tax exemptions, more than twice as Government spending on productive sec- much as in Kenya in relative terms and rough- tors (relative to the total budget) has declined ly half the value of aid. Moreover, since 2005 over the years and reflects the withdrawal there has been a significant increase in the of the state from them such sectors follow- cost of such exemptions. These are given for ing the economic liberalisation of the 1980s. a number of reasons, but only those that aim This trend has begun to change, certainly in to stimulate economic growth are briefly dealt with here. If used as intended, such exemp- tions for productive sector activities increase 71 This figure is surprising because the sector is still relatively small. The authors do not provide detailed evidence for the investments, employment and output growth, claim. eventually leading to higher tax revenues in the 72 Interviews, 1996. long term. Favourable corporation tax rates

45 DIIS WORKING PAPER 2012:06

on profits, deferred tax payments and reduced parastatals could be used to support members import duties are examples of such exemp- of the ruling coalition and help them engage tions. Main beneficiaries are companies es- in business. However, this process took on tablished under the Export Processing Zones many forms, ranging from the outright sale Act, mining companies and other companies of the government’s entire stake in the asset which hold certificates of incentives from the to partial sales, concessions, leases and man- Tanzania Investment Centre and its Zanzibar agement contracts, to the hiving off and sale counterpart (Twaweza, 2010). The extent to of non-core business activities. According which the exemptions promote production to a study by Afrodad (2007), a total of 219 and increase productivity is, however, unclear parastatals were privatized out of 400 marked (for an example, see Mnenwa (2011)). for divesture, and 499 non-core assets were Tax exemptions for the commercial mining sold between 1995 and 2005. Of these 219 sector have also been granted. These have fol- privatized parastatals, Tanzanians came to lowed international standards for the industry own 147 of them through 100 per cent share and appear, therefore, not to have been given sales, 38 were privatized to foreign investors for reasons of patronage. They have ben- through 100 per cent sale, and the remaining efited foreign mining companies, who have 34 were sold through joint venture agree- invested large amounts of capital to excavate ments between the government and Tanza- their mines. However, substantial funds are nian nationals (some of whom were employ- tied up in tax disputes between the mining ees or managers at the respective parastatal). companies and the government (Cooksey and By the end of 2004, the number of parastatal Kelsall 2011). enterprises with central government involve- Looking back in time, the most important ment had been reduced to 47.73 source of rent has probably been related to Based on this information, Khan (2010) the privatisation of parastatals in the after- concludes that privatisation benefited for- math of the SAP crisis. On the one hand, eign investors. Certainly many Tanzanians this privatization did reduce the opportuni- benefited too, but we do not know who they ties for patronage that state ownership of a are, although most are likely to be from the large number of enterprises imply. Mukan- Asian- and Arab-Tanzanian communities. dala (1988) describes how parastatals served Nor do we know what relationship they had as avenues for individual capital accumula- or have with the ruling coalition, or whether tion through the ‘phenomenal haemorrhage’ their businesses have survived. Fischer (2006, of parastatal resources. They were also ‘used 401-407, 555) provides the most recent de- by the ruling coalitions to sort out their intra- tailed information about the privatisation of coalitional struggles…solidify their economic parastatals. According to him it was the party base and ultimately, consolidate their hold on leadership code that held the CCM back from state power.’ Mukanda sees the struggles be- using the privatization process strategically to tween two parts of the ruling coalition, par- further party (members) interests. Moreover, astatal managers and politicians, as ultimate- because only a few party members actually ly destroying the parastatals (quoted from had the funds needed to benefit from priva- McHenry 1994, 148). tization, the CCM could not ensure that its On the other hand, the rents and kick-backs distributed by the subsequent privatisation of 73 Of which eight were at various stages of privatisation.

46 DIIS WORKING PAPER 2012:06

members benefitted significantly from the ways that rents, aid and revenues were used. privatization process.74 What parastatal re- The state slowly lost its ownership of com- trenchees actually did and do for a living is mercial and productive-sector companies and also not well documented. Casual observa- the patronage opportunities that went with tions suggest that only a few of them started them. However, the resource situation has substantial businesses. Some became man- improved considerably since the mid-1990s agers in foreign-owned firms that picked up in terms of both aid and own revenues. Fol- parastatals when these were sold off (Kelsall, lowing the CCM’s legitimacy crisis in the early 2002, 610; Matiku, 2008). Few seem to have 1990s, ruling coalition-controlled resources had the skills, capital and influence needed to are now more clearly used to build and main- buy up privatised companies. And some may tain the coalition and to win elections for the have gone into real estate, hotels, construc- party so it can stay in power. These require- tion or petty business. A indigenous entrepre- ments have brought business-people willing neurial class with close relations to the ruling to contribute money into closer relations with coalition is not emerging from the privatisa- the ruling coalition. In exchange, such people tion process, unlike what seems to have been have benefited from various rents. A larger the case in Mozambique (Pitcher 2011). degree of mutual dependence and interests has emerged.

5.6 Conclusions There is no clear indication that rents and 6. POLITICAL STABILITY the privatisation of state-owned enterprises WITHOUT SUSTAINED have been deliberately allocated to enhance ECONOMIC DEVELOPMENT productive capacity or party interest, while the importance of aid is a striking feature of The country’s political settlement has been Tanzania’s public finances. This does give do- remarkably stable. Thus, the general config- nors considerable influence in various policy uration of the factions making up the ruling arenas, notably to do with macro-economic coalition – the CCM, the bureaucracy and stability and various public-sector reforms, the military – has not changed significantly but in practice domestic political factors exert since the 1960s, although important crises a dominant influence over policy and its im- have caused some internal reconfigurations plementation, especially at the sector level. (in particular the mutiny in 1964, the eco- Since the mutiny crisis of 1964, the mili- nomic crisis of the 1980s and the CCM’s tary has enjoyed relatively generous access to legitimacy crisis, leading to political liber- budgetary resources and has been allowed to alisation in the 1990s). It is only recently engage in commercial activities, although the that private-sector groups and actors have magnitude of these is not well known. The gained influence, although productive-sec- SAP crisis reduced the resources controlled tor interests in (smallholder) agriculture and by the ruling coalition, and also changed the manufacturing, with some exceptions in the latter, are weak. 74 This contrasts with the situation in other countries, such Thus, over time the Tanzanian ruling as Mozambique, where privatization was used strategically by the party to promote its members and interests (Pitcher coalition has moved slowly from a ‘(weak) 2011). dominant party’ towards ‘competitive cli-

47 DIIS WORKING PAPER 2012:06

entelism’, where lower level factions of the has become more difficult to establish pockets ruling coalition have gained strength over of administrative effectiveness at local govern- time. However, the competition for power ment level or to shift state resources downwards and rents mainly takes place between factions that could help to build capacity in implement- within the ruling coalition rather than between ing priority activities in productive sectors such the ruling coalition and excluded factions, as agriculture because it has become more dif- as is normally the case in this form of cli- ficult to control the efficient use of rents that entelism (Khan 2010, 64). These key features many factions lay claim to. Although the ruling of Tanzania’s political settlement help to ex- elite (and, to some extent, donors) are able to plain why the ruling elite has not been able set the policy agenda, it has much less control to implement state initiatives that have led to over the implementation process. This inability significant structural changes in the economy. to discipline rent recipients is a key explana- Nor have the resources controlled by the rul- tion for why state initiatives to promote pro- ing coalition been used systematically to build ductive-sector capacity have been relatively up the capacity of productive sectors. unsuccessful in Tanzania, despite its political Moreover, the reintroduction of com- stability (see also Gray 2011, 268-269; Cooksey petitive elections since 1995 has not reduced and Kelsall (2011)). clientelistic politics. On the contrary, it has Finally, a viable political settlement between motivated the ruling elite to pursue populist the CCM and the largely Asian-Tanzanian rather than growth-enhancing policies. Kjaer entrepreneurs, which dominates the produc- and Therkildsen (2012) have explained some tive sectors in the formal economy (together of the casual mechanisms that motivate rul- with foreign firms), has not yet emerged due ing elites to take such decisions to win elec- to the ruling party’s economic nationalism. tions. Moreover, elections have aggravated This type of nationalism, which has broad the internal fragmentation of the ruling coali- public backing (Peter 2002) and dates back to tion and increased competition among its fac- Nyerere’s attempt at ‘socialist’ development, tions. The political tensions generated by the has hampered the design and implementation union with Zanzibar and the growing threats of industrial policies based on mutual state– to the CCM’s hegemony on the islands fol- business interests, which is a key to sustain- lowing political liberalization are especially able growth and poverty alleviation in poor conflict-ridden. countries. At the same time, the lower level factions Nevertheless, ruling coalition relations of the ruling coalition have gained in strength with capitalist entrepreneurs have gradually following the reintroduction of competitive become more important since the 1980s, for elections in 1995.75 This is partly due to their ruling elites increasingly depend on entrepre- importance in nominating party members for neurs and other people with money to finance political office and mobilising voters for the their political activities to stay in power. The ruling coalition, and partly to the devolution individuals and groups who matter here are, (albeit feeble) of administrative and political in particular, traders and transporters, real-es- power to local governments. Consequently, it tate tycoons, entrepreneurs in the construc- tion business and increasingly politicians us-

75 Although central–local tensions existed much earlier ing their position to accumulate legal and/or (McHenry 1983). illegal wealth.

48 DIIS WORKING PAPER 2012:06

It belongs to the picture that smallholders is that political and economic advantage (the in agriculture, on which most Tanzanians two are increasingly difficult to separate) is depend for a living, are not part of the rul- converted into educational advantage, which ing coalition, nor are significant organised in turn leads to positional advantage, espe- groups of entrepreneurs in the productive cially in the public sector and the political sectors. They do not control substantial and parties. This trend may have accelerated in organised economic power, which could buy recent years but has also changed somewhat. influence in the ruling coalition. This is a re- Circumstantial evidence suggests that many sult of – and reflects – the fact that structural children of top officials in the CCM and changes in the economy have been modest the government increasingly go to private and gradual and that productivity increases schools and subsequently go abroad for uni- in the productive sectors have been limited versity-level education. Upon return these or stagnant. children seek work in the private rather than Should a stronger entrepreneurial group in the public sector or in politics, where op- emerge, this would not, in the foreseeable portunities are now fewer and more insecure future, include many black Africans. Local than in the past. This may also indicate that entrepreneurs of Asian or Arab origin would private-sector opportunities are improving certainly dominate, as they have done for and are gradually becoming more attractive. decades. However, this group is not likely to But some observers argue that for some of engage in the types of business–state rela- the children of the political-administrative tionships that are typically needed to address elite this could also be a possible first step longer term issues of economic and industry towards a later political career launched on policies. Moreover, because a black African the basis of wealth they have earned for entrepreneurial group would probably have themselves in the private sector.76 On the greater political legitimacy than the Indian- other hand, this group may also eventually Tanzanian entrepreneurs, it would also pose a form the nucleus of a well-educated and po- potential political threat to the present ruling litically well-connected group of indigenous coalition. This may also explain why the rul- entrepreneurs of a sort that Tanzania has ing coalition around the CCM seems to have not yet had. fewer problems with foreign-owned com- The attempt over the last couple of dec- panies: they are unlikely to pose significant ades by the ruling elite of the CCM to use threats to the ruling coalition, even though capitalism to ensure the survival of its ‘so- foreign companies (especially in mining) are cialist’ regime’ has so far been successful. occasionally the target of popular protests. But after fifty years of independence and of The way the ruling coalition and the oppo- relative political stability, the foundations for sition parties manage the contradictions be- a strong economy that could lead to sustain- tween economic and political liberalization able poverty alleviation are still feeble. At the in this context is therefore crucial for social same time, the importance of money in poli- stability, as well as in building an indigenous tics has increased with the need to finance the domestic entrepreneurial group that can drive survival of the present ruling coalition in the economic growth and employment. context of competitive clientelism. Another trend, which may have a longer term impact on Tanzania’s political economy, 76 Interviews, November 2010.

49 DIIS WORKING PAPER 2012:06

REFERENCES

Afrobarometer and REPOA. 2008. Key findings on Tanzania. Dar es Salaam: Afrobarometer and REPOA. Afrodad. 2007. Tanzania’s Experience with Privatisation Policies: A Case Study. Harare: African Forum and Network on Debt and Development. Aminzade, Ronald 2003. “From race to citizenship: the indigenization debate in post-socialist Tanzania.” Studies in Comparative International Development. 38 (1):43-63. Anti-Corruption Resource Centre. 2009. Overview of Corruption in Tanzania. Bergen: Transparency International and Christian Michelsen Institute. Arkadie, Brain Van, and Do Duc Dinh. 2004. Economic Reform in Tanzania and Vietnam: A Comparative Commentary. The William Davidson Institute, University of Michigan Business School. Babeiya, Edwin 2011. “Electoral Corruption and the Politics of Elections Financing in Tanzania.” Journal of Politics and Law. 4 (2):91-103. Bana, Benson, and Willy McCourt. 2006. “Institutions and governance: public staff management in Tanzania.” Public Administration and Development. 26:395-403. Banerjee, Abhijit V., and Esther Duflo. 2007. What is middle class about the middle classes around the world? MIT Department of Economics Working Paper 07- 29. Department of Economics MIT Massachusetts. Basedau, Matthias, and Alexander Stroh. 2008. Measuring Party Institutionalization in Developing Countries: A New Research Instrument Applied to 28 African Political Parties. Hamburg: GIGA. Bourgouin, France. 2011. The Politics of Large-Scale Mining in Tanzania: between domestic policy and global economic processes. Mimeo. Copenhagen: DIIS. Cameron, Greg 2002. “Zanzibar’s turbulent transition.” Review of African Political Economy. 29 (92):313-330. Chachage, C. S. L. 1995. “The Meek shall Inherit the Earth but not the Mining Rights: the Mining Industry and Accumulation in Tanzania “ In Liberalised development in Tanzania: Studies on accumulation processes and local institutions, edited by Peter Gibbon. Uppsala: Scandinavian Institute of African Studies. Cimoli, Mario, Giovanni Dosi, and Joseph Stiglitz. 2009. “The political economy of capabilities accumulation: The past and future of policies for industrial development.” In Industrial policy and development: the political economy of capabilities accumulation edited by Giovanni Dosi, Mario Cimoli and Joseph Stiglitz. Oxford: Oxford University Press. Cooksey, Brian. 2010a. The Investment and Business Environment for Export Horticulture in Northern Tanzania. African Power and Politics, APP. London: ODI. ——. 2010b. Predation, rent-seeking and the Tanzanian military. Mimeo. Dar es Salaam.

50 DIIS WORKING PAPER 2012:06

——. 2011. The investment and business environment for gold exploration and mining in Tanzania. Background paper. Africa Power and Politics. London: ODI. Cooksey, Brian, and Tim Kelsall. 2011. The political economy of the investment climate in Tanzania. Research report No 1. African Power and Politics, APP. London: ODI. Cooksey, Brian, John Shao, and Rogasian Mkende. 2007. Cashewnut production, marketing and input supply in Tanzania. Situational analysis 2007. Dar es Salaam: Tanzania Development Research Group. Coulson, Andrew. 1982. Tanzania: a political economy. Oxford: Claredon Press. Donge, Jan Kees van, and Athumani J. Liviga. 1986. “Tanzanian political culture and the cabinet.” Journal of Modern African Studies. 24 (4):619-639. Easterly, William. 2001. “The Middle Class Consensus and economic development.” Journal of Economic Growth. 6 (4):317-335. Erdmann, Gero. 2007. The Cleavage Model, Ethnicity and Voter Alignment in Africa: Conceptual and Methodological Problems Revisited. Hamburg: German Institute of Global and Area Studies (GIGA). Ewald, Jonas. 2002. Electionalism or Democracy? The interface between economic reforms and democratisation in Tanzania. In 1st annual network conference: Actors and Approaches in Local Politics. Oslo. ——. 2011. Challenges for the democratisation process in Tanzania. Moving towards consolidation 50 years after independence? PhD Thesis. Gothenburg: Peace and Development Research, School of Global Studies. Fischer, Pius. 2006. Rent-seeking, institutions and reform in Africa: Theory and empirical evidence from Tanzania. PhD Thesis. University of Konstanz. Frot, Emmanuel, and Javier Santiso. 2010. Crushed aid: fragmentation in sectoral aid. Working Paper No 284. Paris: OECD. Gibbon, Peter. 2005. Options for Interaction between BSPS II and Tanzanian Exporter Associations. Copenhagen: Ministry of Foreign Affairs. Gibbon, Peter 2001 “Civil society, locality and globalization in rural Tanzania: a forty-year perspective “ Development and Change. 32:819-844. Gould, Jeremy, and Julia Ojanen. 2003. Merging in the circle: The Politics of Tanzania’s Poverty Reduction Strategy. Helsinki: Institute of Development Studies, University of Helsinki Gray, Hazel. 2011. Tanzania and Vietnam: A Comparative Political Economy of Economic Transition. PhD Thesis. London: Department of Economics. School of Oriental and African Studies. University of London. Heilman, Bruce, and John Lucas. 1997. “A Social Movement for African Capitalism? A Comparison of Business Associations in Two African Cities.” African Studies Review. 40 (2):141-171. Holtom, Duncan. 2007. “The challenge of consensus building:Tanzania’s PRSP 1998-2001.” Journal of Modern African Studies. 45 (2):233–251.

51 DIIS WORKING PAPER 2012:06

Hoogeveen, Johannes, and Remidius Ruhinduka. 2009. Poverty reduction in Tanzania since 2001. Good intentions, few results. Dar es Salaam: Research and Analysis Working Group. Hvidt, Marianne. 2011. Getting the Story Right – A Growth Diagnostic for Tanzania. Master’s Thesis, University of Copenhagen, Copenhagen. Hyden, Goran. 2005. Why do things happen the way they do? a power analysis of Tanzania. Report. Dar es Salaam: Embassy of Sweden. Iliffe, John. 1979. A modern history of Tanganyika. Cambridge: Cambridge University Press. ILO. 1999. Social and labour Issues in Small-Scale Mines. Geneva: ILO IMF. 2009. Fourth Review Under the Policy Support Instrument— Staff Report; Press Release on the Executive Board Discussion; and Statement by the Executive Director for United Republic of Tanzania. Washington: IMF. Iversen, Torben. 2006. “Capitalism and democracy.” In The Oxford Handbook of political economy, edited by Barry R. Wingast and Donald A. Wittman. Oxford: Oxford University Press. Kelsall, Tim. 2002. “Shop windows and smoke-filled rooms: Governance and the re-politization of Tanzania.” Journal of Modern African Studies. 40 (4): 597- 619. ——. 2003. “Governance, democracy and recent political struggles in Mainland Tanzania “ Commonwealth & Comparative Politics. 41 (2):55-82. Khan, Mushtaq, and Hazel Gray. 2006. State weakness in developing countries and strategies of institutional reform – Operational Implications for Anti- Corruption Policy and A case-study of Tanzania. London: DFiD. Khan, Mushtaq H. 2005. “Markets, states and democracy: Patron-client networks and the case for democracy in developing countries.” Democratization no. 12 (5):704-724. ——. 2010. Political Settlements and the Governance of Growth-Enhancing Institutions. London: SOAS. Kiragu, Kithinji, R.S. Mukandala, and Denyse Morin. 2004. “Reform pay policy: techniques, sequencing and politics.” In Building state capacity in Africa: new approaches, emerging lessons, edited by Brian Levy and Sahr J. Kpundeh. Washington: The World Bank. Kironde, J.M. Lussuga. 2009. Improving Land Sector Governance in Africa: The Case of Tanzania. In Land Governance in support of the MDGs: Responding to New Challenges” Washington DC Kjaer, Anne Mette, and O. Therkildsen. 2012. “Elections and landmark decisions in Tanzanian and Uganda.” Democratization (forthcoming). Lawson, Andrew, and Lise Rakner. 2005. Understanding patterns of accountability in Tanzania: synthesis report OPM, Chr. Michelsen Institute & REPOA. Lerise, Fred S. 2005. Politics in land and water use management: study in Kilimanjaro, Tanzania. Dar es Salaam: Mkuki na Nyota Publishers.

52 DIIS WORKING PAPER 2012:06

Lindemann, Stephan. 2010. Civilian control of the military in Tanzania and Zambia: explaining persistent exceptionalism London: Crisis States Research Centre, London School of Economics. Lindemann, Stephan, and James Putzel. 2008. State Resilience in Tanzania – Draft Analytical Narrative. London: Crisis State Research Center. London School of Economics. Liviga, Athumani. 2009a. Economic and political liberalisation in Tanzania: the unintended consequences. Dar es Salaam: Department of Political Science and Public Administration. ——. 2009b. “Tanzania: a bumpy road to consolidated democracy.” Eastern Africa Social Science Research Review. 25:1-42. Lofchie, Michael F., and Thomas Callaghy. 1995. Diversity in the Tanzanian business community. Its implications for economic growth. USAID mission, Dar es Salaam. Maghimbi, Sam , Razack B. Lokina, and Mathew A. Senga. 2011. The Agrarian question in Tanzania? A State of the Art Paper. Dar es Salaam and Uppsala: Nordic Africa Institute in coperation with the University of Dar es Salaam. Matiku, Emmanuel James. 2008. Evolution of Public Enterprises Reform Policy in Tanzania: Privatisation of Banking, Electricity and Telecommunications. PhD Thesis. Griffith University, Department of Politics and Public Policy. Mattes, Robert. 2008. The material and political bases of lived poverty in Africa: insights from the Afrobarometer. Working Paper No 98. Ann Arbor: Afrobarometer. Mbelle, A. V. Y , L Rutasitra, Beatrice Mkenda, R. Lokina, A. Naho, and Y. Aikaeli. 2010. Analytical study on the drivers of growth and implications for growth stategy for (mainland) Tanzania Dar es Salaam: Department of Economics, University of Dar es Salaam. Mbilinyi, Majorie, and Timothy Nyoni. 2000. Agricultural and livestock policy, 1997. Dar es Salaam: Rural Food Security Policy and Development Group. Institute of Development Studies. McGowan, Patrick J., and Patrick Bolland. 1971. The political and social elite of Tanzania: An analysis of social background factors. Syracuse: Maxwell School of Citizenship and Public Affairs. McHenry, D. E. Jr. 1983. “A measure of harmony/disharmony in a one-party state : Low-level party leaders’ choices for members of Parliament compared with those of both high-level party officials and the people in Tanzania, 1965-1975.” The Journal of Developing Areas 17 (1983). 337-348. McHenry, Dean E. 1994. Limited choices: the political struggle for socialism in Tanzania. Boulder: Lynne Rienner Publishers. McMillan, Margaret and Dani Rodrik. 2011. Globalization, structural change and productivity growth. Tufts University and Harvard Kennedy School. Mercer, Claire 1999. “Reconceptualizing state-society relations in Tanzania: are NGOs ‘making a difference’?” Area. 31 (3):247-258.

53 DIIS WORKING PAPER 2012:06

Mesquita, Bruce Bueno de , and Alastair Smith. 2009. Tanzania’s Economic and Political Performance: A District-Level Test of Selectorate Theory. New York,: Department of Political Science, New York University. Mesquita, Bruce Bueno de, Alastair Smith, Randolph M. Siverson, and James D. Morrow. 2005. The logic of political survival. Cambridge, Massachusetts: The MIT Press. Mmuya, Max. 1998. Tanzania. Political reform in eclipse. Crisis and cleavage in political parties. Dar es Salaam: Friedrich Ebert Stiftung. ——. 2009. Elites and the new poverty agenda. An argument from selectorate theory and empirical evidence. EPP workshop paper. Dar es Salaam: Department of Political Science and Administration, University of Dar es Salaam. Mmuya, Max, and Arne Tostensen. 2007. Tanzania Power and Change Analysis. Dar es Salaam: Dutch Embassy. Mnenwa, Raymond. 2011. Elite coalitions: a factor in the formulation and implementation of the Export Processing Zones in Tanzania. Dar es Salaam: Research on Poverty Alleviation Morley, Louise, Fiona Leach, and Rosemary Lugg. 2009. “Democratising higher education in Ghana and Tanzania: opportunity structures and social inequalities “ international Journal of Educational Development. 29:56-64. MTIM/NBS. 2008. Annual Survey Of Industrial Production 2008. Statistical Report. Dar es Salaam: Ministry of Industry, Trade and Marketing. National Bureau of Statistics. Mukandala, R.S., Kithinji Kiragu, Edriss Maruva, and Benson Bana. 2008. Strengthening the Policy Development Process Dar es Salaam: President’s Office Public Service Management & The Cabinet Secretariat, Office of the Chief Secretary. Mukandala, Rwekaza S., Samuel Mushi, Joel Barkan, and Godwin Njema. 2005. The political economy of Tanzania. Washington: World Bank. Mukandala, Rwekaza Sympho. 1988. The political economy of parastatal enterprise in Tanzania. PhD Thesis. Ann Arbor: University of Michigan. Mwase, Nkunde, and Benno Ndulu. 2008. “Tanzania: explaining four decades of episodic growth.” In The political economy of economic growth in Africa, 1960- 2000, edited by Benno Ndulu, Stephen A. O’Connell, Jean-Paul Azam, Robert Bates, Augustin K. Fosu, Jan Willem Gunning and Dominique Njinkeu. Cambridge: Cambridge University Press. National Bureau of Statistics. 2007. Revised national accounts estimates for Tanzania mainland. Dar es Salaam: National Bureau of Statistics. Nord, Roger, Yuri Sobolev, David Dunn, Alejandro Hajdenberg, Niko Hobdari, Samar Maziad, and Stéphane Roudet. 2009. Tanzania: the story of an African transition. Washington: IMF. Nyerere, Julius. 1969. Freedom and unity. London: Oxford University Press. OECD. 2007. Aid effectiveness: 2006 survey of monitoring the Paris Declaration. Overview of the results. Paris: OECD.

54 DIIS WORKING PAPER 2012:06

Peter, Chris Maina. 2002. Racism, racial discrimination, xenophobia and related intolerances in Tanzania. In The East African Conference on Racism and Xenophobia Zanzibar: Faculty of Law, University of Dar es Salaam. Pitcher, Anne. 2011. Party politics and economic reform in democratic Africa. Cambridge: Cambridge University Press. Ponte, Stefano. 2002. Farmers & markets in Tanzania: How policy reforms affect rural livelihoods in Africa. Oxford: James Curry. REDET. 2010. The practicability of separating business from politics in Tanzania. Dar es Salaam: Redet. Research and Analysis Working Group. 2009. Lost in transition: income poverty reduction since 2001. Dar es Salaam: REPOA. Rweyemamu, Denis. 2009. Strategies for Growth and Poverty Reduction: Has Tanzania’s Second PRSP Influenced Implementation? DIIS Working Paper, EPP, 2009: 13. Copenhagen: Danish Institute for International Studies. Sandbrook, Richard. 2000. Closing the circle. Democratization and development in Africa. London: Zed Books Ltd. Schiavo-Campo, Salvatore, Giulio de Tommaso, and Amitabha Mukherjee. 1997. Government employment and pay in global perspective: a selective synthesis of international facts, policies and experience. Washington: Public Sector Management and Information Technology Team. The World Bank. Shivji, Issa G. 1975. Class struggles in Tanzania. Dar es Salaam: Tanzania Publishing House. ——. 2009. Accumulation in an African periphery: a theoretical framework. Dar es Salaam: Mkuki na Nyota Publishers. Skarstein, Rune. 2010. “Economic Liberalization and Smallholder Productivity in Tanzania. From Promised Success to Real Failure, 1985–1998.” In Tanzania. From Ujamaa to Mkapa, edited by K. J. Havnevik and Aida C. Isinika. Dar es Salaam: Mkuki na Nyota Publishers. Skarstein, Rune, and Samuel Wangwe. 1986. Industrial development in Tanzania: some selected issues. Uppsala: Scandinavian Institute of African Studies. Smith, Benjamin. 2005. “Life of the party. The Origins of Regime Breakdown and Persistence under Single-Party Rule.” World Politics. 57:421-451. Sokomani, Andile. 2005. “Money in Southern African politics.” African Security Review. 14 (4):81-90. Southall, Roger. 2006. “Troubled visionary: Nyerere as a former president.” In Legacies of power. Leadership change and former presidents in African politics, edited by Roger Southall and Henning Melber. Uppsala: The Nordic Africa Institute. Sulle, Emmanuel, and Fred Nelson. 2009. Bio-fuels, land access and rural livelihoods in Tanzania. London: IIED. Sundet, Geir. 2004. The politics of land in Tanzania. Mimeo. Dar es Salaam Temu, Andrew E., and Jean M. Due. 2000. “The business environment in Tanzania after socialism: Challenges of reforming banks, parastatals, taxation and the civil service.” Journal of Modern African Studies 38: 683-712.

55 DIIS WORKING PAPER 2012:06

Therkildsen, O., and Torsten Geelan. 2009. Local council election data 2005 Copenhagen: DIIS. Therkildsen, Ole. 2000a. “Contextual issues in decentralisation of primary education in rural Tanzania.” International Journal of Educational Development 20:407-421. ——. 2000b. “Public sector reform in a poor, aid-dependent country, Tanzania.” Public Administration and Development 20:61-71. ——. 2001. Efficiency, accountability and implementation: public sector reform in East and Southern Africa. Geneva: UNRISD. Therkildsen, Ole 2011. Policy making and implementation in agriculture: Tanzania’s push for irrigated rice. DIIS Working Paper, EPP. 2011: 26. Copenhagen: DIIS. Therkildsen, Ole, and Per Tidemand. 2007. Staff management and organizational performance in Tanzania and Uganda: public servant perspectives. Copenhagen: Danish Institute for International Studies. Tidemand, Per, and Jamal Msani. 2008. Local Government Reforms in Tanzania and their impact on Local Governance and Service Delivery: Empirical Evidence of Trends in Tanzania Mainland 2000 – 2008 Dar es Salaam: Repoa. Tidemand, Per, and Nazar Sola. 2010. Decentralization Reforms in Tanzania. Dege Consultant. Dar Es Salaam: USAID. Tordoff, William. 1967. Government and politics in Tanzania : a collection of essays covering the period from September 1960 to July 1966. Nairobi: East African Publishing House. Treichel, Volker 2005. Tanzania’s Growth Process and Success in Reducing Poverty. Working Paper. Washington: IMF. Tucker, Stevens, Barak D. Hoffman, R.S. Mukandala, and Mark Billera. 2010. Democracy and governance assessment of Tanzania. Dar es Salaam: USAID. Twaweza. 2009. Reforming allowances: A win-win approach to improved service delivery, higher salaries for civil servants and saving money. Dar es Salaam: Twaweza. ——. 2010. Broken Promises in Primary Education: When enough money doesn’t reach schools on time. Dar es Salaam: Twaweza. ——. 2011. Who returned? Performance in the Bunge and MP re-election. Dar es Salaam: Twaweza. UNESCO. 2009. Global Education Digest 2009: Comparing Education Statistics Across the World. Paris: Unesco. United Republic of Tanzania. 2005. Environmental statistics, Tanzania Mainland Dar es Salaam: National Bureau of Statistics. Utz, Robert. 2008. Sustaining and Sharing Economic Growth in Tanzania Washington: The International Bank for Reconstruction and Development / The World Bank.

56 DIIS WORKING PAPER 2012:06

Valentine, Theodore R. 2006. Reforming pay in the Tanzania public service: successes, limitations and challenges. In ESRF Policy Dialogue Seminar. van de Laar, Aart J.M. 1972a. “Foreign business and capital exports from developing countries: the Tanzanian experience.” In Socialism in Tanzania. Vol I, edited by Lionel Cliffe and John S. Saul. Nairobi: East African Publishing House. ——. 1972b. “Growth and income distribution in Tanzania since independence.” In Socialism in Tanzania. Vol I, edited by Lionel Cliffe and John S. Saul. Nairobi: East African Publishing House. ——. 1973. “Toward a manpower development strategy in Tanzania.” In Socialism in Tanzania. Vol II, edited by Lionel Cliffe and John S. Saul. Nairobi: East African Publishing House. Wangwe, Samuel 2010. The political economy of Tanzania: the evolution of policy making. Mimeo. Dar Es Salaam: REPOA. Wangwe, Samuel M., and Brian Van Arkadie. 2000. Overcoming constraints on Tanzanian Growth: policy challenges facing the third phase government. Tanzania Political Economy Series, 2. Dar es Salaam: Mkuki Na Nyota Publishers. Weinstein, Laura. 2011. “The Politics of Government Expenditures in Tanzania, 1999–2007 “ African Studies Review. 54 (1):33-57. White, Howard , and Geske Dijkstra. 2003. Programme aid and development : beyond conditionality London Routledge. Whitfield, Lindsay , and Ole Therkildsen. 2011. What drives states to support the development of productive sectors? Strategies ruling elites pursue for political survival and their policy implications. DIIS Working Paper, EPP. 2011: 15. Copenhagen: Danish Institute for International Studies. Wolter, Denise. 2008. Tanzania. The challenge of moving from subsistence to profit. Paris: OECD- Development Centre. World Bank. 2002. Constructing knowledge societies. Washington: World Bank. ——. 2005. World Development Report 2005. A better investment climate for everyone. Washington: The World Bank. ——. 2007. Tanzania. Sustaining and sharing economic growth. Country Economic Memorandum and Poverty Assessment. Vol II: Main Report. Dar es Salaam: PREM, Africa Region, the World Bank. Wuyts, Marc. 1998. Informal economy, wage goods and the changing patterns of accumulation under structural adjustment. Theoretical reflections based on the Tanzanian experience. ESRF. Yu, Bingxin, and Alejandro Nin-Pratt. 2011. Agricultural Productivity and Policies in Sub-Saharan Africa. IFPRI Discussion Paper 01150. Washington: International Food Policy Research Institute.

57 DIIS WORKING PAPER 2012:06

58